冷藏飲品為現今飲料業銷售主力之一，而便利商店通路更是飲品銷售的主要管道，因此本研究將利用便利商店之POS系統取得個案公司冷藏飲品的實際銷售紀錄，再結合個案公司所提供的訂貨數量，及透過紀錄並觀察影響冷藏飲品銷售關鍵因素之天氣狀況等因素，來建立一套銷售預測模式，以解決個案公司冷藏飲品生產及庫存調度的產銷平衡問題，以避免產品缺貨而影響主力產品在市場的佔有率，且能有效降低產品庫存並降低營運成本。本研究以中部地區某知名食品公司為研究對象（以下簡稱個案公司），並以個案公司兩大冷藏主力產品（產品A、產品B）之主要銷售通路7-11及全家便利商店為研究標的，進行資料分析與整理，並透過SPSS線性迴歸分析建立預測模式及進行預測誤差分析，並將模式實際應用到新年度進行銷售預測，協助相關產銷決策的進行。未來將與現有的資訊系統做結合，利用Web-service線上直接取得未來三天的天氣預測資訊自動進行銷售預測，以利後續以銷售預測為基礎進行生產安排與庫存調度，以期降低整體營運成本。模式預測誤差平均約在13%以內，已明顯提高預測之準確度，未來將可確實提高相關產銷決策的品質。

Refrigerated beverage is the main products for drink industries. CVS is the main sales channel. The study will focus the records for refrigerated beverage sales in CVS and combine with the real order of DC for the CVS all around Taiwan. The main key factors for the sales of refrigerated beverage is the weather condition. We will try to analyze the relation between the sales and the weather condition. For improving the efficiency of stock management and the shortage of goods, a forecasting model will be made based on the relationship for refrigerated beverage sale in CVS. This study is supported by a famous food Co. in middle Taiwan (namely A Co.). We will focus the two main product of the refrigerated beverage. Data analysis and forecasting model will be made. Two main CVS channels for the refrigerated beverage will be included for the forecasting model. The new model will be used for forecasting the sales for new years and the bias analysis will be done for model adjusted. The model will combined with the information system of A Co. The web-service technology will be used for capturing three further days and sales forecasting and bias analysis will be made automatically. The stock management and production scheduling will be done based on the forecasting output for decreasing total cost. The average bias is less than 13%. It's better than the forecasting by managers' experience. The forecasting model will be helpful for the decision of production and distribution.