Accurate academic grade forecasting is crucial for improving student outcomes and optimizing resource allocation in educational settings, but traditional methods often lack complexity. In this paper, we explore using various algorithms for predictive modeling to forecast academic grades. The primary goal is to identify the most effective algorithm for predicting student performance, which can assist educators and policymakers in making informed decisions. Algorithms such as Linear Regression (LR), Decision Trees (DT), Random Forest (RF), and Gradient Boosting Regression (GBR) were utilized to build the models. Each was assessed using metrics including training and testing scores, Mean Squared Error (MSE), Mean Absolute Error (MAE), Median Squared Error (MedSE), and R-squared (R²).The findings revealed that Gradient Boosting Regression (GBR) achieved the highest accuracy, with a training score of 0.999, a testing score of 0.96, an MSE of 0.5068, an MAE of 0.258, and an R² of 0.995. These results indicate that GBR surpasses the other models in predicting academic performance, offering a reliable tool for grade forecasting and supporting educational planning and intervention strategies.