Distributed Data sources integration is a method aiming to offer access for users to multiple databases through requests on a global schema. The use of an ontology is seamlessly the most promising approach for a possible automation to deal
with the semantic heterogeneity, considered as the most important difficulty.
In this paper we propose approach uses a domain ontology to solve semantic conflicts considered as a global schema and a description schema base in order to
manage the mappings, done manually by the administrator of every relational source, between the global schema and the different databases. Moreover, we
describe the algorithms related to relevant sources localization, requests decomposition, inter-sources joint, requests rewriting and recomposition of the results as tuples. Finally, we present a prototype for relational databases integration.
