1960年代起資訊科技應用興起以協助組織運行，多數企業因缺乏資訊知識背景，紛紛購入套裝軟體協助業務營運。但套裝軟體無法切合企業的流程，且隨環境變遷和科技演進，不敷使用的問題日益嚴重。從資料庫設計的角度出發，套裝軟體複雜的資料架構、長期修改和存取資料而欠缺管理、無關連式資料庫的概念，導致組織的資料品質低落。當今組織如何將資料庫重新設計以符合所需、新舊系統資料該如何轉換以提升品質，是企業面臨的一大挑戰。 　　有鑑於此，本研究設計一套資料庫重構流程，以企業流程為基礎為企業設計客製化的資料庫，並將資料從套裝軟體移轉至該理想的資料庫。流程分三階段，階段1是運用資料庫反向工程（Database Reverse Engineering）的方法，還原企業現行資料庫的資料語意和模型架構；階段2則結合流程模型（Process Model）和資料模型（Data Model）的概念，建立以企業流程為基礎的理想資料庫；階段3利用ETL（Extract、Transform、Load）和資料整合的技術，將企業資料從現行資料庫中萃取、轉換和載入至理想資料庫，便完成資料庫重構的作業。 　　本研究亦將資料庫重構流程實做於個案公司，探討企業早期導入之套裝軟體和以流程為基礎的理想資料模型間的設計落差。實做分析結果，二者在資料庫架構設計、資料語意建立和正規化設計等三部分存有落差設計，因此在執行資料庫重構之資料移轉解決落差時，需釐清來源端資料的含糊語意、考量目的端資料的一致性和參考完整性、以及清潔錯誤的來源資料。 　　最後，總結目前企業老舊資料庫普遍面臨資料庫架構複雜、無法吻合作業流程所需、未制訂完善資料庫管理機制等問題，而本研究之資料庫重構流程的設計概念，能為企業建立以流程為導向的理想資料庫。 / The raising of information technique helped organization governance greatly was started since 1960s, but because of lack information background and knowledge, many organizations just simply brought software packages to assist business processes or organization governance. The result was those software packages which couldn't fit in with the processes of organization' requirements were getting worse because of changes of environment. From the view of database design, it results in low quality of data because of the complexity of database structure, long-term modifications and accessing to data, and the lack of relational database knowledge. Nowadays, the problems of redesign database structure or transform data from a old system to a new system are great challenges to enterprises. Based on the above, thie research designed a process of database restruction in order to establish customized database based on businesss processes. There are three phases of this process. In phase 1, a company acquires the original data structure and semantic of its software package by the method of database reverse engineering. In phase 2, using concepts of process model and data model, the company establishes its ideal database based on businesss processes. In phase 3, it extracts, transforms, and load data from the current database of software package to ideal database by the technique of ETL and data integration. After these three phases, the company completes the process of data restriction. The process of database restruction is done in a case company to analyze the design gap between the current data model of software package and the ideal data model based on business processes. In the result of analysis, this research found out there are three gaps between its as-is and to-be data models. These three gaps are the design of database struction, the definition of data semantic, and the design of database normalization. Because of these design gaps, when removing gaps by data transformation, a company should pay attention to clarify the semantic of source data, considerate the consistency and referential integrity of destination data, and clean dirty data from source database. Finanlly, the summary of the problems a company using old database are the complexity of database structure, the unfit database for businesss processes, the lack of database management, etc. The process of database restruction this research design can assist a company in establishing ideal database based on business processes.