Abstract: In order to get tomato and green pepper leaf area accurately and conveniently, scanner was used to get leaf images and AutoCAD was used to get actual leaves area. Regression analysis were carried on between actual leaf areas and leaf length, breadth, product of length-breadth measured in the field respectively, and regression models were established between leaf length, leaf breadth, product of leaf length-breadth and the actual leaves area respectively. The results showed that relations between the leaf length, the leaf breadth and the leaf area all assumes the power function, and relation between the product of leaf length-breadth and the leaf area assumes the linear. Errors analysis was carried on for three kinds of estimating models, and the results showed that accuracy of simulated value were bad for regression models of leaf length and leaf breadth. However, the estimated errors were much lower and precision was highest for the product of leaf length-breadth regression models for two kinds of crop, actual size of leaf area can be reflected truly, and the reduction coefficients were 0.6393 and 0.6509 respectively for tomato and the green pepper.