This study presents a machine learning–based framework designed to predict students’ academic performance in elective courses and the resulting impact on other core courses within the same semester. Using data collected from the Department of Geography and Environmental Science, Begum Rokeya University, Rangpur (BRUR), various models—including Logistic Regression, Decision Tree, Support Vector Classification, Naïve Bayes, and Random Forest Classification—were trained and evaluated. The dataset underwent preprocessing, encoding, and correlation analysis using Pearson, Spearman, and Kendall methods to identify inter-course relationships. Among all models, the Random Forest Classification algorithm achieved the highest performance with 78% accuracy, an R² score of 0.80, and a MAPE of 22%, effectively capturing nonlinear dependencies between core and elective results. The model successfully predicted both elective course performance and the influence of elective selection on other courses of the same semester. These findings demonstrate the potential of machine learning for data-driven academic advising, enabling universities to recommend electives that align with students’ academic strengths. The framework lays the foundation for developing an automated elective course recommendation system to enhance student performance and decision-making in higher education.