Loan approval prediction is among the main areas of financial institution's decision-making processes, facilitating the assessment of individual's creditworthiness and risk management. Despite the influx of loan applications received daily, not all applicants are granted approval, leading to inconsistent decisions and increased default risks. The paper considers different machine learning classification methods, which include Random Forest, LGBM, Gradient Boosting, K-Nearest Neighbor and Adaboost to forecast loan approval. By using cross-fold validation and hyper-parameter tuning techniques, the study strives to enhance prediction accuracy. Evaluation measures such as Accuracy, F1 score, Precision, and Recall are used for the performance of classifier. The results highlight the Random Forest classifier's superiority, achieving an impressive accuracy of 96.15% when employing a k value of 15. Overall, this study contributes to advancing loan approval prediction methodologies by providing insights into effective machine learning techniques for enhancing decision-making processes in financial institutions.