In the contemporary global context, education stands as a crucial pillar of society, prompting governments worldwide to exert substantial efforts towards its improvement. The disruptive impact of the 2020 coronavirus outbreak on daily life has significantly affected educational systems, leading to a documented decline in student performance. This underscores the urgent need for a more concerted and serious approach to address the issue and identify effective solutions, considering the various influencing factors at play. Particularly in these challenging times, innovative strategies are essential to enhance the standard of education and mitigate the escalating failure rates. This study proposes a model geared towards enhancing student performance in exams by predicting success through classification. Employing various classifiers such as Logistic Regression, KNN, and SVM, we aim to determine the most accurate classifier while avoiding overfitting and underfitting through simple techniques. Recognizing diverse factors like family issues and alcohol consumption influencing student performance, our machine learning-based approach seeks to achieve three objectives: predicting exam success, identifying the optimal classifier, and elucidating the most impactful factors. This proactive intervention approach empowers teachers and parents to address challenges before students reach the critical exam stage, facilitating timely problem resolution.