针对数据挖掘中的分类问题，本文提出了一种利用神经网络抽取分类规则都方法。为了易于抽取规则，采用遗传算法对神经网络的结构进行了进化。实际运行结果表明了该方法的有效性。-Classification is an inportant problem in data mining.This paper presents an approach to discover classification rules by using neural networks.Genetic algorithms is used to evolve neural network structure for extracting rules more easily.The experimental result shows the effectiveness of the method.