Existing credit card fraud detection studies are generally based on imbalanced datasets, where various sampling methods are used to alleviate the imbalance before applying machine learning techniques for detection. This study applies machine learning methods to a balanced credit card dataset to evaluate their reliability on imbalanced datasets. Using Kaggle’s balanced dataset of 568,630 transactions, we compared Random Forest, Neural Networks, Logistic Regression and Naive Bayes. Random Forest and Neural Networks achieved near-perfect accuracy (99.9%), demonstrating superior fraud detection capabilities that could significantly enhance financial institutions’ security systems. These findings highlight the significant potential of these advanced machine learning methods to optimize fraud detection systems in financial institutions and provide strong support to further improve fraud detection capabilities.