A machine learning model using a decision tree is proposed to evaluate the credibility of clients applying for loans in Indian banks. It compares decision tree machine learning models studied by other authors to see which one is best for predicting loan approval outcomes, finding that the one discussed in this paper performed slightly better. To the best of our knowledge, existing research has focused on general loan applications, with no prior studies specifically addressing education loans. Our study fills this gap, achieving an accuracy score of approximately 94% in accurately classifying (approved and not approved) education loan application approval. We drew inspiration from the loan application forms of various Indian banks to inform our attribute selection. A key innovation of our approach is that it enables predictions about loan approval without relying on credit scores, which is particularly relevant for education loan applicants who often lack established credit histories.