Prunus mume was studied as research object in this paper. Image segmentation, feature extraction, classification, etc. were studied based on the image of P. mume. A P. mume species recognition system was automatically established based on these technologies. Different methods were proposed for describing the color, shape and texture features of P. mume respectively. With its color feature extraction, the method suitable for describing the color characteristics of P. mume image was proposed through the results of a histogram classification. Its flatness describing its shaped feature was improved. Only the flower region was calculated for its flatness whereas the background region was not included. The algorithm was simpler than the original method and the final result was not to be affected using the improved method. The method of calculating gray level concurrence matrix (GLCM) was improved. The flowers smallest external rectangular region was extracted firstly as the new length and width of the image before calculating GLCM. Then the matrixes were calculated for four angles. The corresponding matrix elements were cumulated and divided with the results by 4 The average matrix was then used for the new GLCM, and the related parameters of the new matrix were calculated as part of texture features. At last, the SVM (support vector machine) was used for classification and recognition. Experiments were conducted for 660 P. mume images (60 images each for 11 different types). Three hundred and thirty images were used for the training samples, and the rest of 330 images was used for the test samples. After cross validation being conducted for SVM classifier, the experimental results showed that the average recognition accuracy reached 93.94%. The system has high recognition accuracy and stability. It is important to propagate the P. mumue, reduce the burden of professionals and increase the appreciation of P. mume.