This study investigates the application of Bayesian models in predicting student dropout across multiple classes. Traditional methods often struggle with the complexity of multiclass scenarios and fail to capture uncertainty adequately. Leveraging Bayesian frameworks, our proposed model offers a robust solution by incorporating prior knowledge, handling uncertainty, and providing probabilistic predictions. Through comparative analysis with existing methods, including logistic regression and support vector machines, we demonstrate the superior predictive performance of our Bayesian approach. The findings underscore the efficacy of Bayesian models in multiclass dropout prediction.