In this paper, we propose an Internet public Web service matching approach that paves the way for (semi-)automatic service mashup. We first provide the overview of the solution, which requires a detailed review of two fundamental models - schema/graph matching and semantic space. Based on the conceptual model and the literature study, the complete service matching approach is then provided with four essential steps - semantic space, parameter tree, similarity measures, and WSDL operation matching. The system demonstration that proves the concept proposed in this approach is finally presented. The solution has the potential to facilitate the Internet services mashup.