An apple recognition method with color difference R-G and color difference ratio (R-G)/(G-B) was presented.The apples in the images taken under frontlighting and backlighting conditions were recognized. The contour images were gained through the pretreatments such as noise removal, area filling etc. Shape features were extracted from the contour images based on genetic algorithm (GA). For the clustered or overlapped apples, the method of multiple running GA and converting the contour points of every apple into background in turn was applied. The experiment results showed that the proposed recognition method, to a certain extent, eliminated the influence of shade, backlighting and soil. The recognition rate reached 97%. The features extraction method based on GA segmented the clustered or overlapped apples successfully. Meanwhile, the center points and radiuses were quickly and accurately extracted.