The growing demand for a Semantic Web is focusing attention on Semantic Web trials that use a huge amount of data from a relational database (RDB). Hence, the literature on the transformation of RDB data to resource description framework (RDF) statements has consequently become very important. The relational data model (RDM), the base of RDB, represents real world data using relationships . In existing research, the mapping of RDB data to RDF only generates an RDF triple statement from an RDB table structure in a flat level; existing research does not consider the mapping of RDM which describes the business rules. In RDM, the relationship describes real world business rules by connecting the two entities. If one entity relates to itself, then it has a recursive relationship which helps to explain the logical data model and extend its flexibility. The recursive relationship is a necessary relation type of most data models to describe more complex business rules. This paper suggests a mapping method to map the two types of recursive relationships to the resource description framework schema (RDFS). One type of recursive relationship is generated during the integration process with different entities at different levels in a hierarchical structure while the other is generated at the hierarchical structure between the instances of an attribute in an entity. The transformed RDFS, including the class, subclasses, and subproperties minimizes the loss of data meaning and enables the process of the inference function to be used with RDB data. The method of transforming RDB data to inferable, transitive RDF/RDFS with SQL implies that the most difficult step of writing a Semantic Web can be automated. In short, the proposed method enables a Semantic Web to be built with greater efficiency.