In view of such problems of aircraft in high maneuvering flight as the uncertain aerodynamic parametersso many unknown external interference factors and possible errors in system modelinga nonlinear adaptive backsteeping controller based on RBF neural network was designed.The generalized uncertainty of aircraft in high maneuvering flight was online approximated by RBF neural networkand the weight matrix of neural network was online updated by adaptive law.The “terms explosion” problem in backstepping design caused by repeated derivation of the virtual control law was solved through introducing a first order filter.By construction of Lyapunov functionit was proved that all signals in the closedloop system were bounded and the tracking error was converged to a small neighborhood around zero.High maneuvering flight simulation of some aircraft was carried outand results showed that the designed controller has good tracking effectiveness and robustness.