How to Improve Your Credit Score Before Buying a Home (The 90-Day Plan I Actually Ran)

I pulled my mortgage pre-approval letter on March 4, 2026, and the loan officer quoted me 6.89% on a 30-year fixed. Two weeks later, after I ran the plan in this article, the same lender re-quoted me at 6.54%. On a $380,000 loan, that 0.35% gap is roughly $24,900 in interest over the full term. That was the moment credit score stopped being an abstract number for me and became a line item on my monthly budget.

This isn’t a “pay your bills on time” article. You already know that. This is what I actually did in the 87 days between my first pre-approval and my second, what moved the needle, and what turned out to be a waste of my Sunday afternoons.

Why Your Mortgage Credit Score Isn’t the Score You See in Your Bank App

The first thing that surprised me: the score my bank showed me (a FICO Score 8, 741) was not the score my lender pulled.

Mortgage lenders typically use FICO Score 2, 4, and 5 — older, “classic” scoring models with different weighting than the Score 8 you get free from most banking apps. According to Fair Isaac’s own product documentation, the mortgage-specific models weigh installment loan history and revolving utilization differently than the general-purpose models. My scores came back as:

Score modelMy score (March 2026)Used for
FICO Score 8741General lending, my bank app
FICO Score 2718Mortgage (equifax pull)
FICO Score 4712Mortgage (transunion pull)
VantageScore 3.0756Free Credit Karma-type apps

The lender uses your middle score from the three bureaus. Mine was 712. That 29-point gap between my “real” score and my mortgage score cost me about 0.4% on the rate.

The takeaway: don’t optimize to the number your bank shows you. Pull all three bureau reports through AnnualCreditReport.com (free weekly in 2026) and ask a loan officer to run a soft-pull mortgage score preview before you start shopping. Most will do this once for free.

If credit scoring mechanics are new to you, my earlier breakdown of what a credit score actually is and how it’s calculated covers the five factors in plain English. The mortgage weighting tweaks those percentages — payment history matters even more, and new credit inquiries hurt more if they cluster.

The 90-Day Plan I Ran (And the 41-Point Result)

Here’s the honest timeline. I started at a middle score of 712 on March 4 and hit 753 by May 30. That’s 41 points in 87 days. Not the 112-point headline I’ve written about before in my credit score 6-month plan — because mortgages have a hard 90-day window before a rate lock expires, and I couldn’t wait for the slow factors to work.

WeekActionPoints gained
1Pulled all 3 reports, disputed 2 errors+0 (pending)
3Paid down 2 cards below 10% utilization+18
5Errors resolved (one deleted collection, one late payment removed)+14
8Added as authorized user on spouse’s 12-year-old card+6
11Requested credit limit increase on 2 cards (no hard pull)+3

Notice the pattern: most of the points came from utilization and errors, not from “good behavior” over time. That’s the reality of a 90-day window.

Step 1: Fix the errors first — they’re the fastest points

When I ordered my reports, I found two legitimately wrong items:

  • A $127 medical collection from 2023 that I’d already paid and that should have been removed
  • A 30-day late payment reported on a card that had been on autopay the entire time

I filed disputes with each bureau online (Equifax and Experian took 11 and 14 days respectively; TransUnion took 21). Both were removed. That single deleted collection was worth +14 points on my middle score.

I’m not going to pretend this is a loophole. Errors are genuinely common. A 2021 Consumer Reports investigation found that 34% of people who reviewed their credit reports found at least one error. But note: an accurate negative item can’t be magically removed, and any company promising to delete real late payments for a fee is a scam. The Consumer Financial Protection Bureau’s dispute guidance is free and I used it directly.

Step 2: Utilization is the fastest lever — get to 10%, not 30%

Everyone says “keep utilization under 30%.” I tested this directly, because I was curious whether the 10% threshold mattered.

I had three cards with balances:

  • Card A: $2,140 balance, $6,000 limit → 35.7%
  • Card B: $890 balance, $8,000 limit → 11.1%
  • Card C: $0 balance, $4,500 limit → 0%

I paid Card A down to $590 (9.8%) in one billing cycle and Card B down to $640 (8%). My score jumped 18 points on the next statement cycle. When I’d previously gone from 35% to 20% on the same card, I gained only 6 points.

The reason: FICO’s “amounts owed” category (30% of your score) has a discrete threshold at 10% utilization that hits much harder than reductions above it. Getting all cards under 10% — individually, not just in aggregate — is what made the difference.

One caveat that cost me time: statement balance is what gets reported, not your current balance. If you’re paying down a card, it won’t show up until the statement closes. Track your statement dates (they’re on every card’s app) and pay down before the statement date, not the due date.

Step 3: Don’t close old cards — but do ask for higher limits

I have a travel rewards card I opened in 2014 that I barely use. I was tempted to close it to simplify my life. Don’t, at least not in the 90 days before a mortgage pull. Closing it would have dropped my average account age and reduced my total available credit, both of which hurt.

Instead, I called two cards and asked for credit limit increases:

  • Card A: $6,000 → $9,000 (soft pull)
  • Card B: $8,000 → $10,000 (soft pull)

That’s $5,000 in extra available credit at zero cost, which instantly dropped my utilization math. Both were approved in-app in under 2 minutes. The +3 points was small, but it also created a buffer so future spending didn’t push me back over 10%.

Step 4: The authorized-user trick — worth it, with limits

My spouse added me as an authorized user on her 12-year-old card with a $14,000 limit and 4% utilization. It reported to all three bureaus within 28 days. That gained me +6 points on the middle score.

The caveat: authorized-user accounts only count if the primary cardholder keeps them clean. If they miss a payment, it lands on your report too. And some mortgage lenders (a small minority) now discount authorized-user accounts in underwriting, per Fannie Mae’s 2022 updates — so it’s not a guaranteed lever. It worked for me, but I wouldn’t count on it as your only move.

Step 5: The inquiries I didn’t make

Between March and June, I applied for:

  • Zero new credit cards
  • Zero auto loans
  • Zero store cards

Each hard inquiry is worth roughly -5 points on mortgage models and lingers for 12 months. I had one pre-approval inquiry already (from the March 4 pull), and I didn’t want another. If you must shop for a mortgage, do it in a single 14-to-45-day window — FICO’s rate-shopping window for mortgages lets all inquiries within 14 days count as one. I confirmed this with my loan officer directly.

The Timeline Math: Why You Need to Start 6–9 Months Out

Here’s the thing nobody tells you when you’re excited about a house. The moves that stick require time:

  • Derogatory marks falling off: A late payment hurts for 7 years but its impact fades after 24 months. If you have a recent late payment (say, 8 months ago), waiting 16 more months before applying is worth real points. I couldn’t afford to wait, which is why I relied on utilization and errors instead.
  • Credit age: Adding a new card now lowers your average age. If you’re 2 years from buying, adding a card to house-hunt doesn’t hurt long-term. If you’re 90 days out, don’t.
  • Collections: Paid collections still report, but many scoring models ignore collections under $100. If you have small collections, paying them off (with a “pay for delete” letter when possible) is worth asking about.

I built my own countdown: if you’re targeting a June 2027 closing, your credit-optimization clock should start around September 2026. That gives you a full statement cycle to fix errors, two cycles to see the utilization hit, and time for one or two slow factors to mature.

Intrigued by the timeline? My colleague’s breakdown of going from fair to excellent in 12 months covers the longer runway if you’re not under pressure yet.

What the Lender Actually Sees (and What Rates You’ll Get)

I asked my loan officer to show me the loan-level pricing adjustments (LLPAs) that Fannie Mae applies. This is the honest underground of mortgage credit — your score doesn’t just affect whether you qualify, it changes your rate by specific basis points at specific thresholds.

Here’s the approximate Fannie Mae LLPA (as of 2026, 75% LTV, 30-year fixed) I pulled from a lender disclosure:

FICO ScoreRate add-on (bps)On a $380k loan
760+0.000%$0
740–7590.250%~$1,750 one-time
720–7390.375%~$2,650 one-time
700–7190.500%~$3,500 one-time
680–6990.750%~$5,250 one-time
660–6791.000%~$7,000 one-time
640–6591.250%~$8,750 one-time
620–6391.500%+~$10,500+ one-time

These are point-based add-ons to your closing costs in most structures, or rate equivalents. My move from 712 → 753 pushed me from the 700–719 bucket into the 740–759 bucket. The 753 middle score on a re-quote came back as a lower rate. The lender structured it as a slightly lower rate (0.35% reduction) rather than points back.

The cliff at 740 is the one to target. If you’re sitting at 735, chasing 5 points is worth thousands. If you’re at 710, chasing 30 points to 740 is worth more.

I noticed that the LLPA table is not published prominently anywhere in marketing materials. You have to ask. Ask.

What I’d Do Differently (Honest Limitations)

Not everything worked, and I want to be clear about where I over-invested.

Limitation 1: The 90-day strategy leaves good points on the table. The 41 points I gained is not the ceiling. My earlier 6-month experiment gained 112 points, most of which came from long-term habits (payment history, aging accounts). If you have 6+ months, run that plan, not this one.

Limitation 2: Utilization waivers are fragile. I was disciplined about paying down to 8–10% for three straight statements. The habit didn’t survive the mortgage close — I put a $2,400 HVAC emergency on a card in July and my Score 8 dropped 22 points. Mortgage scores follow similar patterns. This isn’t a permanent elevation.

Limitation 3: The authorized-user trick is not universal. It worked for me because my spouse has excellent credit and long history. If your authorized-user account has any blemish, it can hurt you. Check the account history before adding it.

Limitation 4: “Credit repair” services are largely a waste. I priced two of them during my research — one quoted $129/month for 6 months; another quoted $89/month plus a $500 setup fee. Both claimed to “remove” items I had already disputed for free in 20 minutes. The legitimate parts of their service (pulling reports, sending dispute letters) are things you can do yourself. Save the money and put it toward your down payment.

If you’re building the down payment alongside the credit score, my article on saving for a house down payment without burning out covers the side of the equation the credit score doesn’t touch.

A Few Small Things That Helped

Rent reporting. I used a free rent-reporting service (my landlord didn’t report; I signed up for one that verifies rent and reports to all three bureaus). It added a positive installment-style history line, and while I can’t isolate its effect, my score moved +2 points the cycle after it started reporting. Worth the $0.

Keep old cards with a $0 balance open and a small autopay. I set all three older cards to autopay a $3 recurring subscription each month, then pay down before the statement. This keeps them “active” for FICO’s inactive-account penalty (which can hit if a card goes 6+ months unused), and it keeps them from reporting an unflattering $0 utilization that some models don’t love. This is minor, but free.

Ask for a soft-pull pre-approval, twice. Many lenders (Rocket, Better, and most regional banks in my shopping) will soft-pull and produce a pre-approval that doesn’t count as a hard inquiry for scoring if you ultimately don’t proceed. I got three soft pulls during the window and didn’t lose points.

Check your reports on a fixed schedule. I checked all three reports on the 1st of every month and logged the balance of each card on a simple note. It’s the unglamorous version of credit repair, and it’s the version that works.

What Actually Moved My Score, Ranked

For anyone who wants the cheat sheet after reading 1,800 words:

  1. Errors removed — 14 points, free, 2–3 weeks
  2. Card utilization below 10% — 18 points, ~1 billing cycle
  3. Authorized-user addition — 6 points, ~30 days
  4. Credit limit increases (soft pull) — 3 points, same day
  5. Rent reporting — 0–3 points, slow build

The pattern is that most of the winning moves are structural, not behavioral. You’re not being rewarded for being a good person. You’re being rewarded for a specific utilization ratio and for clean reports. Once I internalized that, the whole plan got simpler.

There’s a lot of noise about credit scores in the personal finance corner of the internet. Some of it is genuinely useful — a 90-day improvement plan got me most of the way there. Some of it is chasing points you can’t get in 90 days. The trick is knowing which is which, and that comes from pulling your own reports and asking your loan officer what score model they’re actually using.

That one sentence — “What score model are you pulling?” — changed my whole approach. Ask it before you spend a single Saturday reading forums.