本文回顧傳統與新發展的離散選擇模式及其在各領域之應用。首先介紹多項羅吉特與多項普羅比模式的型式、特性及校估方法。由於多項羅吉特模式具有方案不相關且獨立特性而限制其適用範圍，因此新發展的離散選擇模式，考量較具彈性的方案變異或共變異結構。另一項重要的發展是離散選擇模式考量個體異質性，反映於方案的變異數、共變異及效用函數參數偏好。離散選擇模式分析顯示性偏好(真實情境下之選擇)及敘述性偏好(假設情境下之選擇)資料，然而每種資料各有其優缺點。整合顯示性及敘述性偏好資料，可以發揮截長補短的功效。離散選擇模式在各領域的研究相當多，但極少應用於國防軍事。本文最後建議未來研究課題。

This paper reviews traditional and recent developed discrete choice models and their applications in a variety of fields. The formulation, property and estimation approach of standard multinomial logit and multinomial probit models are described. Due to the independence from irrelevant alternatives property that restricts the use of the multinomial logit model, advanced discrete choice models have been developed by accommodating more flexible variance or covariance structures of alternatives. Another important extension is the incorporation of individual heterogeneity introduced through variance, covariance and taste parameters. The discrete choice models allow analysis of revealed preference (actual choice) and stated preference (hypothetical choice) data. Each data type has its advantages and limitations. By combining revealed and stated preference data, the strengths of each data type can be achieved while alleviating the weaknesses. While the discrete choice models have been used in a wide variety of disciplines, the application in national defense or military research is very limited. Finally, potential issues for future research are suggested.