With the popularity of Web services, how to discover suitable Web services to support Web services composition has become a challenge. Traditional keyword search is insufficient due to its lower recall and precision. This paper proposes an effective Web service discovery strategy based on Web services description information. In this paper, the description information for Web services is divided into operation information named as WSDL attributes and profile information, where profile information consists of service function information and service constraint information, and they are represented as general attributes and instance attributes respectively. Based on the information and ontology knowledge, three steps are adopted for discovering Web services. Firstly, according to WSDL attribute, the heterogeneities of service operations are resolved. Secondly, based on general attributes, the Web services to meet the function requirement of customers are found. Thirdly, a matching model is applied based on instance attributes and suitable Web services are obtained and provided for customers. The service discovering approach is simple, available and become highly effective by introducing semantics, moreover, it has been applied to e_Scope successfully