近年來，公部門或私部門多有土地價值評估之需要，現今土地交易課稅稅基與私部門投資抵押評估上，多未考量土地與建物個別因素應相互影響價值之因素，導致已開發土地價值評估產生偏誤，故如何合理自房地價格中分離已開發土地價值，是現行土地評價之重要課題。在房地價格分離之相關研究中，如何精準區分土地和建物價值影響因素為其關鍵的基石，而不同房地價格分離方法間差異，亦為本研究探討之議題。 本研究利用特徵價格模型分析方法，以臺北市區分所有建物之住宅交易案例為研究對象，建立複迴歸模型以分離房地價格。實證結果指出臺北市房地價格隨屋齡增加，呈現先遞減後遞增之變化，驗證高屋齡下土地再開發價值之存在，並發現各屋齡分層間土地移轉面積、管理維護對房地價格影響程度不同，藉以有效區分屋齡對土地與建物價值之影響；而現行實務上常見之土地抽取法，在分離房地價格上恐有高估已開發土地價值，以及放大價值漲幅與縮小價值跌幅之問題，顯示其建物成本價值評估過於僵固，將產生偏誤，使建物價值無法有效隨市場景氣變動。此外，透過分量迴歸模型檢驗，其結果顯示已開發土地價值與高屋齡效果間呈現正向關係，亦指價值愈高，其土地再開發價值影響愈顯著。最後，本研究統整相關實證結果，釐清現行房地價格分離上之問題，盼能改善與提升土地價值之評估能力與精準度。 / In recent years, land value assessments have been needed for different purpose. The current assessing of land transaction taxable base and mortgage demands does not consider the individual factors of lands and structures should affect the value of each other, resulting in the developed land value assessment has generated errors. Therefore, how to decompose housing price reasonably and analyze the developed land value is important for land value assessments. 　　This study establishes multiple regression model by selecting residential condominium sale data in Taipei City as sample, in order to decompose the housing price into developed land and structure value. The empirical results show that the reversion of housing price is due to the redevelopment value, and land area and level of maintenance have impact on housing price in different age variables. Besides, we also find that it has a positive relationship between developed land value and redevelopment value effect on price, and land extraction method probably has some problems in overestimation of developed land value, resulting in miscounting of land price fluctuation. According to the aforementioned, this study concludes with some issues of decomposition housing price for improving and enhancing the accuracy of land value assessments.