地名是地圖的要素亦是地圖的語言，對地圖的表現及使用上扮演著極為重要的角色，而其在不同比例尺之地圖簡化作業上，考量清晰易讀及保留重要者等因素，必須要對地名進行合理的選取或刪除，但這個作業，實際上將因為地圖比例尺及所在區域之自然與人文環境之不同，有著極大的岐異度，產生了選取數量和選取對象這兩項棘手的問題，故一個適切的選取模式是非常必要的。 本文以花蓮縣、臺南縣、臺中縣、臺南市、臺中市等五縣市為研究區，針對1/25,000、1/50,000、1/100,000、1/250,000、1/500,000等五種比例尺地形圖上之地名為研究對象。經由實際的地名計測，並利用保留率分析、選取級數分析、迴歸分析、基本字差異分析及地理資訊系統軟體之空間分析功能，將地名選取之原則模式化，解決選取數量和選取對象問題，以提供地圖簡化作業之實務應用。

Geographical name is not only a fundamental element but also a communicating language for maps. It plays a very important role in map usage. While map generation was performed, some factors must be taken into account for the representation of geographical name, for example, easily reading, important-item keeping, etc. consequently, geographic name will be reserving or deleting during the reasonable selection. However, the way of reasonable selection is great variation in different scale and area, so a proper model of the geographical name selection is very important for map generalization. This paper investigated the variations of the geographical name o the series map with scale of 1/25000, 1/50000, 1/100000, 1/250000, and 1/500000 in five counties. We expect to model the principles of geographical name selection. This will be fulfilled by using the methodologies of geographical name investigating, reserve ratio analyzing, selective series analyzing, regression vocabulary characteristics differencing, and the spatial analysis function of GIS. The suitable models will provide the compiler a practical application in map generalization.