This paper focuses on the issue for translating Web metadata, especially Topic Maps and RDF/RDF Schema (RDF). For the Semantic Web, all Web information resources should be accepted and understood by any browsers and web applications. Topic Maps and RDF are both representative international standards for description of Web information resources. However, they have different schemes and syntaxes, and thus much research has been studied to achieve the interoperability between both standards. However, there still remain several issues such as semantic loss, high structural complexity, and so on. This paper proposes an improved method for translating Topic Maps to RDF. Our method translates Topic Maps to RDF based on topics, association roles and occurrences (TRO). The proposed method covers explicit semantics and implicit semantics by the object and the semantic mapping between Topic Maps and RDF. In addition, our method has strength in terms of the completeness and the naturalness. To illustrate the advantages of our proposal, we conduct experiments with several Topic Maps data sets and perform a comparative evaluation between our proposal and the previous methods.