The research of Web Services attracts the attentions of a lot of people in recent years. As a key technique in the development of Web Services, automatic Semantic Web Sevices composition is a problem that must be conquered. Since the similarity between Web Servicec and AI actions, AI planner is the most direct method to resolve this problem. However, due to the different application domain and the constraint of preconditon among AI planners, none of the existing planners can deal with the problem perfectly. In this paper, we present an algorithm for translating Semantic Web Service composition problem to AI planning problem. After the translation, we can choose the most suitable AI planner to compose the Web Services. There are also several key procedures of the translating algorithm in this paper.