One of the extensive procedures in the banking services is the approval of a loan. With the use of Machine learning models and other modern technologies, the loan approval process might be made more efficient and streamlined. One long-standing and critical issue for the banking sector is the prediction of loandefault rates. There is a lot of variances in lending choices and loan defaults happened more often since banks and other financiers utilized subjective criteria and manual processin the past. With the help of ML techniques, banks and other financiers can develop prediction models that are both more accurate and more dependable. This researchuses ML algorithms to extract patterns from a common dataset of loan eligible individuals. Data balancing and exploratory data analysis are done to prepare the dataset. A number of algorithms have been explored including XGBoost, GradientBoosting, and CatBoost. The primary aim of this research is to preprocess data, featureselection, data balancing, split into training and testing sets, classifying data, and comparing performance using classification parametrs such as Accuracy, Precision, Recall and F1-score. The findings showed that CatBoost obtained 92.28% accuracy, GradientBoosting achieved 88.39%, and XGBoost Classifier achieved 86.61%. The obtained results prove that machine learning techniques can have the potential to increase loan approval rates while decreasing default rates.