智慧型資料庫的研究重點，在於發展一個能結合專家系統的知識表達、推論及問題解決能力與資料庫的大量資料儲存管理能力的資訊系統。目前有四種主要類型的結合方式：(1)增強式的資料庫系統、(2)增強式的專家系統、(3)獨立專家系統與資料庫的鬆散結合、(4)專家資料庫。 　　由於目前企業中已有許多資料庫系統在運作之中，上述的結合方式以獨立專家系統與資料庫的鬆散結合方式較為合適，其它三類則因需要對現有系統配合新的架構加以修改，在成本效益上較不經濟。 　　因此，本論文針對獨立專家系統與資料庫的鬆散結合架構進行研究。首先從文獻中瞭解其它相關研究對結合所提出架構的優缺之處，而從資料庫的獨立性、專家系統的獨立性、系統的複雜度、系統的擴充性等四點考慮，提出一個較為獨立、有效率，且建立容易的系統架構。並以財務診斷此一特定應用領域，成功地發展出一套結合財務診斷專家系統與會計資料庫系統的雛形系統，驗證了本架構的可行性。最後並提出進一步的研究建議，以供後續研究參考。 / The reserach objective of the integration, of expert systems and database systems is to develope an information system which can integrate both the capabilities of ex-pert systems (inferencing and problem solving) and capabilities of databases (storing and managing of huge amount of data). There are four ways to integrate them: (1) enhanced databases, (2) enhanced expert systems; (3) loose coupling of independent expert systems and databases, (4) expert database systems. 　　Since there are many existing database systems, the loose coupling of independent expert system, and database is considered to be the most appropriate approach. The other three ways are not cost/benefit feasible because they need major modification of the curresn systems. 　　The focus of this reserach is on loose coupling of independent expert system and database. The strengthes and weaknesses of the relevent researchs are discov-ered through literature review. There are four major considerations in developing the architecture: database independency, expert system independency, system, complexity, system, extendibility. The proposed architecture is more independent, efficient, and easy to be bulit. This research has also built a prototype system coupling a financial diagnosis expert system, with an accounting database. The feasibility of the prototype has verified the usefulness of the proposed architecture. Future reasearch suggestions are also stated.