The precision agriculture for oil palm plantation has received an increasing attention. This circumstance has prompted experts to develop the most accurate oil palm crop production forecast system. This study investigates the utilization of XGBoost for predicting oil palm crop yield planted on mineral and peat land. The annual rainfall and rainy day indices were included as contributed variables. The performance of XGBoost default single train test model was compared to XGBoost Cross Validation and XGBoost grid search tuning technique for hyperparameters configuration. The results showed that the XGBoost model is capable of producing highly accurate predictions, with an R-squared of 87.8332%, MAE of 1.411909 and RMSE of 1.838361. The implemented grid search tuning technique had no discernible effect on the model performance. The feature importance indicates that the previous year's rainfall rate has the greatest influence on the prediction