This project employs decision tree models to identify the most relevant predictors influencing students’ academic performance. The dataset was collected from undergraduate students at the Universidad Mundo Maya Campus Villahermosa. Preprocessing techniques were implemented in RStudio, and the models were trained using RPART using the Gini index to obtain the importance of each variable. This approach highlights the importance of including machine learning methods in this type of analysis, as they enable the discovery of complex patterns and relationships between variables that would not otherwise be easily identifiable. This analysis provides valuable insights into how sociodemographic and academic factors may positively or negatively affect academic performance, offering a solid foundation for future research and the design of strategies aimed at optimizing educational quality and student success.