Abstract: In order to improve the coal and gangue identification rate, a method of coal and gangue texture feature automatic identification was proposed on the basis of the gray level co-occurrence matrix (GLCM). Analyzed the basic principle of the GLCM, characteristic parameters, used GLCM extract of coal and gangue image the angle second moment, correlation, contrast and entropy of the four characteristics as texture features, using support vector machine (SVM) to identify, and the simulation on MATLAB implementation. Research results show that extract texture features with GLCM, SVM recognition method can effectively describe the texture characteristics of coal and gangue, provide important reference basis for the identification and classification of coal and gangue.