At present, in the crop disease harm degree is graded mainly by measure with the eye or paper cut primarily, which is greatly influenced by subjective factors, and results in obvious error. For improvement on identification precision of crop disease, this paper developed a new crop disease grading method based on computer image processing. Image preprocessing, segmentation and statistical calculation were applied effectively in this study. According to crop disease harm degree and classification standard, the crop disease harm degree was determined by computing the proportion of sickness spot area and the normal area on the leaf. The experiment results indicated that the identification accuracy was greatly improved, the grading time and costs is reduced by the manual evaluation, providing accurate data for the study of the crop’s other aspects and has a broad prospect for application.