The problem in this project is about exporter's inventories which is certainly very difficult to prepare the goods that every month. If one provides goods inventory, it will increases the sale of goods. Also the types of goods sold are varied, making it difficult to manage how much inventory space must be prepared for one type of item. For this reason, it requires sales prediction per item in order to estimate inventory of goods that must be prepared. The predictions method used was Linear Regression. In addition, statistical approaches are needed to analyze inventory. With these two methods, it will facilitate decision making for inventory items to be prepared. The final result of this project is to implement a model to predict inventory good that needed to be added and test how accurate the results of predictions are with actual data.