This study investigates the use of data mining techniques to categorize and forecast high school pupils in need of academic support. We create predictive models to identify students at risk of failing final exams by evaluating a variety of data sources, including academic achievement and demographic traits. Data mining techniques in models classify students according to their potential for success or failure. The approach involves collecting and preparing academic data, training machine learning models, and evaluating their accuracy in predicting students who require prior intervention. This predictive method proactively provides academic support to students in difficulty, increasing their chances of success in the baccalaureate exam. The study highlights the effectiveness of using data science tools to target academic interventions, optimize resource allocation, and improve student outcomes.