介绍了人工神经网络的基本概念及在刀具磨损监测中的应用，例举了两种典型网络（ＢＰ网络、ＳＯＭ网络）的结构与学习算法。研究表明，应用前馈神经网络构造的刀具磨损状态二分类器、三分类器及磨损估计网络，检验效果很好。文中提出的ＳＯＭ网络在磨损估计方面具有学习快、精度高的优点，有实际应用价值。 The basic concepts of artificial neural network and its application in tool wear monitoring are introduced. The structures and learning algorithms of two typical networks (BP network and SOM network) are illustrated. The research shows that the test results of the classifier, the classifier and the wear estimation network of cutter wear states constructed by feedforward neural network are good. The proposed SOM network has the advantages of fast learning and high precision in wear estimation, and has practical application value.