This study evaluates the efficacy of three machine learning models—Tuned Random Forest, Multiple Linear Regression, and Tuned Support Vector Regression (SVR)—for regression problems in a Hybrid Microbiology Course. Data was collected from both online and on-site sources, encompassing quizzes, assignments, attendance records, laboratory results, and engagement metrics from platforms such as Google Classroom. The models were assessed using Mean Squared Error (MSE) and R2 Score. The Support Vector Regression (Tuned) model exhibited optimal performance, attaining the lowest MSE (16.85) and the greatest R2 Score (0.88), signifying exceptional predictive accuracy and goodness of fit. The findings indicate that the SVR model is the best appropriate for forecasting student performance, potentially improving tailored learning experiences through the utilization of critical data insights.