Among of reasons for creation of the web ontology is the lack of semantic data stored in relation database (RDB). Ontology may evolve over time, where for some reasons like performance we must keep synchronization between RDB and ontology. This paper proposes an approach to detect any changes applied between two versions of the same ontology, thus updating the RDB schema instead of creating a new schema to reduce redundancy of unchangeable data and reserve a large space. The approach uses a set of SPARQL queries executed on ontology for extract changes between two versions, and then generate SQL queries for lunching in the RDB