Abstract: Control system of multibody satellite is designed to improve attitude stability in this paper. The control law consisting of variable structure control and neural network control is obtained by using Lyapunov stability theory. Variable structure control is a nonlinear control strategy employing feedback of a discontinuous signal, and is very robust to matched plant uncertainties and external disturbances when state variables enter the sliding surface. Neural network systems are physical cellular systems that can acquire, store, and utilize experiential knowledge. The knowledge is in the form of stable states or mappings embedded in networks that can be recalled in response to the presentation of the data input to the network systems. The application of the neural network is made possible by recognizing function approximation capability in this paper. Since dynamics of multibody satellite can not be obtained accurately, the conventional controllers will not ensure the good performance of control system. A suitable technique cope with such system is variable structure control and neural network control. By Integrating their merits, robust capability and high control accuracy can be achieved. In addition, compensation control is introduced into the feedforward loop to eliminate the disturbance caused by the motion of antenna, so convergence of angular velocity is in the form of exponent guaranteed. Simulation results illustrate the control system has the good attitude stable characteristics using the proposed controller in the presence of external disturbances and parameters uncertainties.