针对一类非线性不确定系统 ,提出一种新的基于神经网络动态补偿的最小方差控制方法(MVNNC)。这种控制系统将传统的最小方差控制技术与神经网络优良的非线性逼近能力相结合 ,从而能有效地消除不确定性引起的控制误差。仿真实验表明 ,这种最小方差神经控制系统具有较强的鲁棒性和良好的动态性能 Aiming at a class of nonlinear uncertain systems, a new minimum variance control method based on neural network dynamic compensation (MVNNC) is proposed. This control system combines the traditional minimum variance control technique with the excellent nonlinear approximation ability of neural networks, so that the control error caused by uncertainty can be effectively eliminated. Simulation results show that the minimum variance neural control system has strong robustness and good dynamic performance