Zinc is a commonly used roofing material in Indonesia, alongside roof tiles and similar products. In the production of red sand zinc, two types of raw materials are used: **base coat** (base layer) and **top coat** (top layer). This study aims to analyze and process data on excess raw material usage in the production line, warehouse requests, and production reports to predict raw material needs before production begins. The study applies the **Linear Regression Algorithm** in data mining, as it can generate predictions according to existing patterns and integrate raw material usage with production reports. Testing with RapidMiner produced an RMSE of 0.500, categorized as good. The accuracy, calculated using the MAPE formula, was 0.50%, equivalent to 99.5% accuracy. These results indicate that Linear Regression is effective for predicting raw material usage quickly, accurately, and reliably, enabling companies to plan material requirements efficiently and improve production management.