For the attitude control problem of the variableswept wing aircraft during the deformation process, this paper uses the dynamic inversion theory and the timescale decomposition principle, combined with the characteristics of the neural network to eliminate errors, to design the neural network PID adaptive dynamic inversion controller for the variableswept wing aircraft. Based on Datcom sofware, the relationship model between deformation and aerodynamics of the deformed aircraft is established. In the controller design, the attitude movement of the aircraft is divided into fast and slow systems, and the neural network is used to adaptively adjust the PID parameters to enhance the control ability of the system. The simulation results show that the adaptive control method proposed in this paper can stabilize the attitude angle of the sweptwing aircraft during the deformation process more efficiently than the traditional PID control.