This paper proposes an aerodynamic parameter identification method that combines Long Short-Term Memory (LSTM) and Back Propagation (BP) neural networks. Addressing the issue of lagging effects of angle information on aerodynamic parameters during high angle of attack maneuvers, the method utilizes the advantage of LSTM networks to extract temporal features and obtain the temporal characteristics of angle of attack  \(\alpha \) and sideslip angle  \(\beta \) . Leveraging the powerful nonlinear fitting capability of BP networks through backpropagation, the LSTM-extracted temporal feature vectors, along with other relevant state variables, are input together into the BP neural network for fitting. Simulation results demonstrate that the proposed algorithm achieves high identification accuracy during large angle of attack maneuvers. Furthermore, the network retains good identification performance even when artificial Gaussian white noise is added to the input features.