针对一类不确定离散非线性系统,提出了一种神经网络预测控制算法.考虑系统中的不确定项,建立神经网络辨识模型作为预测模型.为减少重构误差对系统的影响使用了反馈校正技术.为提高控制性能引入了一种动态补偿器来镇定跟踪误差系统.所提出的控制算法保证了闭环系统的所有信号都是有界的.最后,针对AUV中的路径跟踪问题对所提出的控制算法进行仿真应用,仿真结果说明了算法的有效性. For a class of uncertain discrete nonlinear systems, a neural network predictive control algorithm is proposed. Considering the uncertainties in the system, a neural network identification model is established as a predictive model. To reduce the influence of reconstruction error on the system, feedback correction Technology. To improve the control performance, a dynamic compensator is introduced to stabilize the tracking error system. The proposed control algorithm ensures that all the signals in the closed-loop system are bounded. Finally, Control algorithm to simulate the application, simulation results illustrate the effectiveness of the algorithm.