It is difficult to fulfill effective control since nonlinear and uncertain factors for an antenna servo system. In this paper, a compound sliding-mode control strategy is proposed. In order to control servo system effectively, depart it into two parts. One is reference model without considering disturb and uncertain, the other is real model including disturb and uncertain. To one part, choose reasonable sliding surface and design controller law. To another part, RBF neural network is taken in order to acquire upper bound of uncertain. Simulation and experiment results show the effect of control scheme.