文中提出一种模糊规则自动生成方法．该方法借助Ｋ－Ｎｅａｒｅｓｔ－Ｎｅｉｇｈｂｏｒ的概念确定控制曲面的关键点，然后根据关键点确定模糊划分，并由此构造模糊神经网络学习模糊规则．神经网络采用ＢＰ算法学习，在学习过程中可根据收敛情况适当增加模糊分区，并重构神经网络继续学习．该方法能生成较精简的规则集，并具有良好的收敛性和较快的收敛速度 In this paper, a method of automatic generation of fuzzy rules is proposed. The method determines the key points of the control surface by using the concept of K-Nearest-Neighbor, and then determines the fuzzy partition according to the key points, and thus constructs the fuzzy neural network learning fuzzy rules. Neural network uses BP algorithm to learn, in the learning process can be based on the convergence of the appropriate fuzzy partition, and to reconstruct the neural network to continue learning. The method can generate a streamlined rule set and has good convergence and fast convergence speed