Aiming at the problem of maneuver control depending on the standard maneuver library or driver experience, a method based on Monte-Carlo tree search algorithm to generate maneuvering trajectory is proposed. It only needs to give the initial and final state of the target maneuver, and can search through repeated searches. The sequence of operations to achieve the maneuvering target is obtained; and a composite controller with feedforward and feedback is designed to improve the effect of trajectory tracking. The experiment is carried out with the example of the somersault maneuver. The simulation results show that the obtained index of the tendon reference trajectory is close to the best performance of the professional fighter pilot. At the same time, compared with the traditional PID controller, the composite controller can significantly improve the tendon trajectory. The tracking effect provides an effective method to the problem of maneuvering flight control