Ontologybased semantic information integration can resolve the schematic semantic heterogeneity among distributed data sources but can not resolve the context heterogeneity. This situation leads to the semantics of results returned from the information integration system is still unclear. To this deficiency the context representation and inference mechanism is introduced to existed semantic information integration system to extent the capable of resolving semantic heterogeneity. In the extended system, the context heterogeneity can be detected and resolved automaticly so that the complete semantic heterogeneity solution can be formed.