Abstract: Currently the methods of coal-rock recognition have achieved some stage results,but this still cannot meet the production demands. In order to develop a new method of coal-rock recognition,the technique of coal-rock image feature extraction and recognition based on dictionary learning was researched. The dictionary learning algorithm is proposed to apply for extracting the feature of coal-rock image. The method of random selection is adopted to initialize and update the dictionary. The classification algorithm is adopted to identify the coal-rock image. The results show that the feature extraction method by dictionary learning can simply and effectively express the characteristic information of coal-rock image,and achieve a high recognition rate. Research results could provide a new method for the coal-rock in- terface automatic recognition.