Real world expert systems require a large amount of data, which are usually managed by an object-oriented database management system recently. However, a large portion of data still exist in the relational databases. To accommodate such a mixed environment, this thesis study a development of an expert system, DAS-X, which can coordinate multiple relational databases and multiple expert systems with their corresponding object-oriented databases. Since a number of conflicts may arise during the data transmission between relational databases and object-oriented databases due to the differences of schema and expressions on same information. We utilize metadata to review the correspondence. We also develop an Expert System DAS-X, which performs a rule-based updating on replicated data. DAS-X also contains capability of message handling.