Abstract: Because of the lack of coal-rock methods, a novel coal-rock recognition method was proposed based on max-pooling sparse coding in order to explore new coal-rock image recognition methods and efficiently handle high-dimensional coal-rock image data. This method adds the pooling operation when extracting coal-rock image features and adopts the integrated classifier, which consists of multiple weak classifiers when classifying coal-rock images. The experimental results show that this feature-extraction method based on max-pooling sparse coding can simply and effectively express the characteristic information of coal-rock images, greatly enhance the distinguishability of coal-rock images, and achieve a high recognition rate. This method also has good recognition stability. The results obtained herein could provide a new idea and method for automatic coal-rock interface recognition.