The traditional convolutional neural network method can loss some feature information, which may lead tounsatisfied terrain classification accuracy in the field of hyperspectral. In order to solve the problem, a new hyperspectral imagesclassification method based on the 2-D and the 3-D, named hybrid convolutional neural network, was proposed. This methodmainly extracted features from the spatial enhancement aspect and the spectral-spatial aspect. Firstly, a 3-D-2-D convolutionalneural network hybrid structure was proposed for enhance spatial information. Secondly, the 3-D convolutional neural networkstructure was used for joint feature extraction from the aspect of spectral-spatial, and then the spectral-spatial comprehensiveseparability information was obtained. Finally, the separately obtained information was feature fused and classified. This methodwas used for classification experiments on hyperspectral data sets and compared with other methods. The results show that theclassification accuracy of this method is 99. 36% and 99. 95% respectively in Indian Pines and Pavia University data set, and itsclassification accuracy and kappa coefficient are also better than other methods. This method has a competitive advantage in theclassification of hyperspectral images.features; feature extraction