Abstract: The aim of the Web service is to connect all the services provided by the different organizations，different companies，and even different persons seamless and to build the loosely-coupled Web application to make resources sharing maximum.The key question is how to find and choose the suitable services quickly and effectively.In order to decrease the amount of the services finally participated in the matchmaking process so as to improve the discovery efficiency，the matching algorithm do some preprocesses on the service set.And this paper also proposes an approach to compute the similarity between semantic web services based on the specific features of web services，proposes the idea of infinite classification.At last the paper evaluates the performance in practice.

摘要： Web服务的目的是为了将不同组织、公司甚至个人提供的服务进行无缝的连接，建立一种松耦合的Web应用，实现资源共享的最大化。而在众多候选服务集中快速、准确地发现和选择所需要的服务即服务匹配成为其中关键的问题。在服务匹配过程中，为了减少最终参与匹配的服务的数量，提高服务发现的效率，匹配算法对参与匹配的服务集进行了预处理。并且依据服务描述本身的特点，对服务匹配中基于语义的Web服务相似度计算方法进行了研究，提出了无限分类的思想。最后实现了匹配算法并对性能进行了对比。