Abstract A half-scale aircraft vertical tail model is taken for research object in the paper. To solve uncertainty and nonlinear problem of the model structure vibration control system，the paper presents a thorough research on modeling the system using neural network. The single-frequency signal and different sweep frequency signals are used as excitation signal respectively. Based on improved nonlinear auto-regressive moving average model and double BP neural network of delayed feedback, system identification and dynamic modeling of the vertical tail model structure vibration control system are carried out. The results show that the methods have high identification precision and short training time，the neural network model has good generalization ability, and the identification results are reasonable and reliable.