In this paper a simplified control system framework is presented to solve maneuvering trajectory tracking problem of uninhabited air vehicles. The flight control system is divided into three subsystems including command generation, transformation and allocation. According to the kinematics equations of the aircraft, flight path angle commands can be generated by desired 3D position from path planning. These commands are transformed to body angular rates through direct nonlinear mapping, which is simpler than common multi-loop method based on time scale separation assumption. Then, by using weighted pseudo-inverse method, the control surface deflections are allocated to follow body angular rates from the previous step. Additionally, an on-line multilayer neural network is used to provide robustness to the uncertainty of system. Numerical simulation results show the viability of this approach.