Which BMR formula is most accurate?

Mifflin-St Jeor is the most accurate of the common predictive equations for the general population, validated in studies to be within about 10% of measured resting energy more often than Harris-Benedict or Owen.

Why is Harris-Benedict still used?

It is the oldest and most familiar equation (1919, revised 1984) and is built into many older tools and textbooks. It still works reasonably, but it tends to overestimate BMR by a small margin compared with Mifflin-St Jeor.

Do I need body fat percentage for an accurate BMR?

Only if you use the Katch-McArdle equation, which is based on lean body mass and can be more accurate for very lean or very muscular people. For everyone else, Mifflin-St Jeor using weight, height, age and sex is sufficient.

Mifflin-St Jeor vs Harris-Benedict for BMR

If you have used more than one online calorie calculator you may have noticed they disagree. Often that is because they use different BMR equations. The two most common are Mifflin-St Jeor and Harris-Benedict. Here is how they differ and which to trust.

The answer first

Use Mifflin-St Jeor. It was derived in 1990 from a modern population and is the most accurate predictive BMR equation for most people, usually landing within about 10% of measured resting energy. The older Harris-Benedict equation (1919, revised by Roza and Shizgal in 1984) is still reasonable but tends to read slightly high. Our TDEE & BMR calculator uses Mifflin-St Jeor.

The formulas

Both equations have the same structure — a weight term, a height term and an age term — but different constants.

Mifflin-St Jeor (1990)
  men:   BMR = 10·kg + 6.25·cm − 5·age + 5
  women: BMR = 10·kg + 6.25·cm − 5·age − 161

Harris-Benedict (Roza & Shizgal 1984 revision)
  men:   BMR = 88.362 + 13.397·kg + 4.799·cm − 5.677·age
  women: BMR = 447.593 + 9.247·kg + 3.098·cm − 4.330·age

A worked comparison

Take a 35-year-old woman, 65 kg, 168 cm:

Equation	Calculation	BMR
Mifflin-St Jeor	10·65 + 6.25·168 − 5·35 − 161	1364 kcal
Harris-Benedict	447.593 + 9.247·65 + 3.098·168 − 4.330·35	1418 kcal

Harris-Benedict reads about 54 kcal (≈4%) higher here. The gap is usually small but it is consistent — Harris-Benedict tends to overestimate, especially in people who are overweight, because it was built on a leaner early-20th-century sample.

What the evidence says

When researchers compared predictive equations against BMR measured by indirect calorimetry, Mifflin-St Jeor predicted resting energy within 10% of the measured value more often than Harris-Benedict or the Owen equation. That is why most modern dietetics guidance defaults to it.

When to use something else

Katch-McArdle uses lean body mass instead of total weight, so it can be more accurate if you are very lean or very muscular — but you need a body-fat estimate first. Get one from the body-fat calculator.
Any equation is just a starting point. Whichever you pick, multiply by an activity factor to get TDEE and then adjust from real-world weight trends, as explained in how to calculate your TDEE.

Bottom line

For the overwhelming majority of people, Mifflin-St Jeor is the best default. The differences between equations are smaller than the day-to-day noise in your weight, so do not agonise over which one to use — pick Mifflin-St Jeor, get a number, and let your actual results fine-tune it.

This is general fitness information, not medical or dietary advice. See our methodology for sources.