In view of the loop-type flight process of military aircraft, the computational model of maneuvering flight was established by taking split-s and immelman as examples. According to the minimum maneuvering speed setting, the speed constraint condition was set. The average maneuvering segment of the bucket was divided. By using the fitting function of the neural network, the non-linear relationship between the maneuvering speed and the maneuvering condition was obtained. By using the global optimization function of genetic algorithm and combining the weights and bias results obtained by neural network, the overload selection method for maximizing boundary flight performance in maneuvering process was explored. The simulation results show that the overload control law flight obtained by genetic algorithm optimization can make the aircraft have a larger altitude-velocity range than the conventional control method given in the relevant manual. At the same time, the method in this paper is faster than that of combining genetic algorithm with numerical integration directly. The proposed method can also be used in the optimization design of flight trajectories of other types of aircraft loop-type and has certain practical value.