A significant concern in universities is student dropout, which reflects the institution's quality and affects its reputation and the students' future careers. High dropout rates can directly impact a private university's budget, marketing strategy, and sustainability. Therefore, it is crucial to improve the quality of higher education, reduce student dropout rates, and enhance academic planning. Analyzing dropout patterns is essential for effective management. This study aims to predict dropout rates of university student in Bangladesh by using various machine learning techniques. In order to create a machine learning model for forecasting the dropout rate of students, we used seven well-known classifiers: SVM, KNN, XGBoost, Decision Tree, Logistic Regression, Naïve Bayes, and Random Forest. The dataset included around 500 Bangladeshi university students who answered 29 questions. The collection comprises records with different properties. After preprocessing, the data were split into two sets: one for training and the other for testing. The classifiers were subsequently trained and tested. To identify the most suitable classifier for this problem, we thoroughly assessed the performance of all seven classifiers. Our findings indicate that XGBoost and Random Forest have 98.06% accuracy rate which is highly promising for predicting dropout rates. By utilizing machine learning to predict university dropout rates, early intervention, more effective resource allocation, and enhanced student support are made possible, which in turn improves retention rates, academic results and can be useful for developing the education system in the university of Bangladesh.