生物多樣性對於人類和地球其他生物甚為重要。當缺乏物種多樣性，面對自然災害和氣候變遷時，生態系統就變得脆弱。隨著物種滅絕率增加，我們在兼顧科學知識和人類需求的基礎上，選擇適當和正確的方法保護生態系統和稀有物種。隨著電腦軟硬體技術迅速發展，使得自動化的系統和跨平台操作更具穩定性和實用性。以GIS為平台，結合遙測和GPS實地調查的資料，使物種分布模擬廣泛應用於生態領域。本研究以海拔、坡度、坡向、坡面位置和植生指標為預測變數，應用六種演算法建立物種分布模式，預測瀕危稀有的蘇鐵蕨之潛在分布，並探討六者之用途及適用性。此六者於蘇鐵蕨分布的推測表現皆在水準之上，高低排序為抉擇樹、最大熵值法、BIOLCIM、廣義線性模型，而DOMAIN及區別分析殿後。六者於空間外推表現欠佳，乃因地形變數之侷限與樣本不足，有待將來找出更有效之預測變數。背景主體樣本比在三至八倍為適當但未必最佳，但仍需驗證其通用性。東距及北距位置因子確有助於提升稀有物種推測能力，但不利空間外推，故應謹慎用於廣泛分布物種。六者可大幅縮減後續現場調查面積，節省所需龐大資源，並藉迭代模擬改善推測結果。六模式中的前四名可協助劃設生態保護區，而後二名可協助尋找稀有物種新族群，各具用途。除此之外，也能將其分布透過GIS展示，加強民眾對稀有物種生態保育之認識，達到環境教育之目的。 Biodiversity is very important for humans and all other species on the Earth. Without the diversity of species, ecosystems are more fragile to natural disasters and climate change. With an increase in the rate of species extinction, we must conserve ecosystems and rare species by choosing right methods that are sustainable on the basis of appropriate science and human needs to conserve ecosystems and rare species. Species distribution modeling (SDM) uses 3S technology and statistics and becomes increasingly important in ecology. The study was intended to predict the potential habitat of Brainea insignis (cycad-fern, CF) in central Taiwan by using six statistical methods coupled with GIS techniques. The spatial distribution of the species was examined by overlaying the layer of CF samples collected with GPS on the layers of elevation, slope, aspect, terrain position, and vegetation index derived from SPOT images. The six models were developed and validated with different data sets. The performance of DT was the best, followed by maximum entropy (MAXENT), BIOCLIM, generalized linear models (GLM), DOMAIN, and discriminant analysis (DA) was the worst. They had poor performance on spatial extrapolation because of the limitations of topographic variables and insufficient CF samples, and thus more effective predictor variables will be searched out in the future. Ratio of background to target samples falling within 3–8 was appropriate but not necessarily optimal, and its general applicability will need further confirmation. Location factor of easting and northing coordinates greatly improved the predictive ability of models for estimating rare species, but was not good for extrapolation. It should be used carefully for wide-ranging species. The six models remarkably reduced the area of subsequent field surveys, thereby saving large amounts of resources and improving outputs by iterative modeling process. The first four of six models may help delineate conservation areas and the last two may aid in new population discovery of rare species. They serve their own purpose individually. In addition, the distribution of rare species from SDM may be displayed in a GIS to strengthen public knowledge of ecological conservation to attain the purpose of environmental education.