Data integration based on ontology has become an effective method to cope with the heterogeneous data on the Internet. In this paper, we consider the problem of data integration and use an ontology-based framework to address this problem at a semantic level, we apply the idea of ontology as a tool for data integration. An improved data integration method is proposed by analyzing the existing data integration system. This method uses ontology web language to represent the global ontology, local ontology and the mapping rules between them. The mapping rules can be modified according to the dynamic change of data source. The method has three main stages: building the shared vocabulary, building local ontologies and defining mappings.