UD. Udang Makmur is a shrimp farming business that often faces challenges in accurately predicting sales stock due to reliance on manual forecasting methods. This study aims to develop a web-based sales prediction application utilizing the AutoRegressive Integrated Moving Average (ARIMA) method. The application uses daily sales data from January to December 2023 for analysis. The results indicate that the ARIMA (2,1,1) model delivers accurate predictions, achieving a Mean Squared Error (MSE) of 0.264295. Forecasts for the next 24 periods demonstrate a stable projection, with predicted values converging around 2.5 and a narrow 95% confidence interval. These findings highlight the model's reliability and low uncertainty for the forecasted time frame. The application was successfully tested using the Black-Box method, confirming its functionality and effectiveness in supporting sales predictions.