针对一类未知非线性动态系统 ,提出了一种基于神经网络的自适应输出跟踪控制方案。网络权值的自适应修正规则是基于Lyapunov稳定性理论实现的 ,避免了递归训练过程 ;一个滑模控制项用于消除神经网络逼近误差的影响。因此 ,该自适应神经控制器能保证系统的全局稳定性和输出跟踪误差渐近收敛于 0。 Aiming at a class of unknown nonlinear dynamic systems, an adaptive output tracking control scheme based on neural network is proposed. The adaptive correction rules of network weights are based on Lyapunov stability theory, avoiding the recursive training process. A sliding mode control item is used to eliminate the influence of the approximation error of neural network. Therefore, the adaptive neural controller can ensure that the global stability of the system and the output tracking error converge asymptotically to zero.