Aiming at the problem of military aircraft flight maneuver recognition, combined with the characteristics of flight data, a deep learning model ConvLSTM is proposed to realize flight maneuver recognition. Obtain the flight data of a certain type of aircraft, and preprocess the original data. According to the characteristics of the flight data collected by the flight parameter recording system of the aircraft, the corresponding ConvLSTM model is constructed. The model is trained and verified by experiments. The experimental results show that the algorithm has a high recognition rate for military aircraft flight maneuver recognition, and the recognition accuracy of seven types of flight maneuvers reaches 91.5%, which lays the foundation for the evaluation of flight control quality.