In the present era, the industrial 4.0 revolution marks technological advancements in the fisheries sector, making fish farming industries increasingly effective and efficient. Predictions related to future fish harvest production can be used as considerations in planning. This research aims to build a prediction system for fish farming harvests by comparing four regression methods: linear regression, polynomial regression, random forest regression, and support vector regression to obtain accurate predictions. With the existence of this harvest prediction system, it is expected to provide information about the quantity of future harvests, enabling harvest partners to take appropriate steps to enhance their profitability. Based on the research results, the best regression model is polynomial regression. This model yields an average Root Mean Square Error (RMSE) value of 4.39 and an average Mean Absolute Percentage Error (MAPE) value of 0.41%. This indicates that the regression model has good and accurate prediction capabilities.