本文系利用灰色系統理論建立臺灣地區鳳梨零售價格預測模式，分別為GM(1,1)滾動式GM(1,1)及GM(1,N)三種灰色系統模型。灰色預測模式不若多元迴歸模式需要大樣本及較好的分布規律等條偉，而是將原始數據經灰色生成轉化為生成數，並以灰微分方程組（式）構建為基礎，得到生成數具有永續效果之意義。研究結果發現，三種預測模型皆有良好的預測準確度，而其中以經濟理論為基礎所構建之GM(1,N)有較佳預測能力，每人每年平均消費量與國民所得為影響零售價格之主要變數。

　　　　　The purpose of this study is to apply the grey theory to develop a model for predicting thepineapple retail price in Taiwan Area. Three grey system models, GM(1,1) the recursive of GM(1,1)and a variety of polyfactor forecasting model of GM(1,N) are presented.GM differs from multiple regression which need large amount of data with conventionalstatistical distribution. Accumulated generating operation (AGO) functions are the essentialconcepts and the basis of grey differential variables.The results show that average quantity consumed per capita per year and national income arefound to be important factors influencing the retail price. By using the grey system model, thepredicted pineapple retail price are very accurate and effective.