本文利用神经网络聚类学习方法对机械故障的故障模式进行识别分类，通过实验研究，证明该方法作为一种新的自适应模式识别技术，比传统的聚类方法和基于ＢＰ神经网络故障模式识别方法具有较高的模式分类能力。 In this paper, neural network clustering learning method is used to identify and classify the failure modes of mechanical faults. Experimental studies show that this method is a new adaptive pattern recognition technology, which is more efficient than traditional clustering methods and BP neural network based fault pattern recognition The method has high pattern classification ability.