Aiming at the traditional methods of text classification, the dimensions need to be reduced, the features are extracted manually, and the classification accuracy is poor, furthermore, convolutional neural network CNN can only extract local information, cannot better express context information, long short-term memory network LSTM can extract context dependencies, and the classification effect is good, but the training time is long, a text classification algorithm based on hybrid CNN-LSTM hybrid model is proposed. The algorithm uses the Skip-Gram (continuous skip-gram) model and the CBOW (continuous bag-of-words) model in word2vec to represent words as vector, using CNN to extract local features of text, LSTM saves historical information, extracts contextual dependencies of text, and uses the feature vector output by CNN as the input of LSTM, using Softmax classifier for classification. Tests on the Chinese news corpus of Sogou.com show that the algorithm can effectively improve the precision of text classification.