With the development of machine learning, this technology is more and more used in people’s daily life, such as object classification, portrait recognition, text analysis and so on. In recent years, researchers have tried to apply machine learning to Earth Sciences, such as image recognition of geological ores. In this project, we apply the method of deep learning to the classification and recognition mineral, and propose a classification algorithm based on deep learning. we obtain the key attribute of classification, namely feature selection, through accurate data, and add a classification algorithm on this basis, so that the category can be obtained quickly when new samples are encountered. At the same time, our experiments are based on type of minerals dataset.