With the large-scale development of electric vehicles, the number of public charging piles in operation and the charging capacity are increasing year by year. However, there are many problems in the operation of charging pile, such as frequent failures, difficult operation and high maintenance costs, and traditional fault detection methods are inefficient. Therefore, a hybrid network prediction method of electric vehicle charging pile operation state based on convolution neural network (CNN) and long short term memory (LSTM) network is proposed, which can realize the comprehensive evaluation of the operation state of electric vehicle charging piles. In the feature data input stage, the key indicators of the operation state of the charging pile are analyzed, and the characteristic quantities of the influencing factors of the operation state are extracted through CNN. Then, the operation state of the charging pile is judged and predicted by LSTM, so as to realize the early warning of potential failures of the charging pile. The experimental results show that the method has high prediction accuracy and strong practicability. It can accurately reflect and predict the operation state of charging piles, and can be used in actual charging pile fault prediction and operation and maintenance.