A neural network-based dynamic inverse control approach to a missile is proposed in this paper. The missile system is separated into two loops according to time-scale separation principle. Taking account of the fast loop and the relatively slow loop, the dynamic inversion controllers are designed in both of the two loops for decoupling control, and neural networks-based controller is designed for the slow loop to compensate the uncertainty and disturbances of the whole system. The RBF neural networks are used to approximate the slow states model of missile, and the approximation errors of the ...