Surface areas of fruits and vegetables are often estimated by assuming that they resemble a shape from which surface area can be mathematically calculated. Two such models, sphere and ellipsoid, were compared to three alternative approaches for estimating surface areas of apple fruit(Malus domesticaBorkh.) of four cultivars ('Royal Gala’, ‘Braeburn’, ‘Red Delicious’, and ‘Granny Smith'). Sphere and ellipsoid models underestimated actual surface area by 15 and 18%, respectively, with poor correlation between predicted surface area and actual surface area (estimated by measuring the area of adhesive tape required to cover the fruit surface). A method based on finite elements marginally overestimated surface area but reliably predicted surface area of small ‘Royal Gala’ and ‘Braeburn’ apples and had reasonably close correlation with actual surface area. Non‐linear regression models, developed from the very strong correlation between actual surface area and both fruit mass and volume, were the most accurate predictors of fruit surface area of the methods tested.