In the paper,an anomaly detection model of user power data based on LSTM hybrid algorithm is proposed based on the study of CNN and LSTM.Firstly,a detector for feature extraction of power data is proposed.The detector extracts abstract features from samples,reconstructs the input,and compares the reconstruction error with the threshold to realize the high-dimensional feature representation of power users.Secondly,this paper proposes a hybrid cnn-lstm power data anomaly analysis network,which uses CNN superposition feature representation,and captures the context of power data based on double-layer LSTM,so as to effectively improve the classification and regression ability of the model.In the experimental stage,taking the power data provided by a power company as an example,the proposed model is verified.The results show that the performance of the proposed model is the best,and the accuracy and recall are 89.3% and 69% respectively.