The detection of credit card fraud remains a critical challenge for the financial sector requiring for innovative solutions to effectively manage potential risks. This work examines the effective use of supervised machine learning approaches, specifically Random Forest (RF) and Artificial Neural Networks (ANN), which have been found to be powerful tools for fraud detection. Using a dataset of transactional characteristics, we show using these techniques, achieving excellent accuracy and reliability for identifying fraudulent activities. Both of these methods have their own benefits especially with respect to imbalanced data and recognized fraud patterns making them valuable candidates for this domain. By presenting their implementation and evaluating their performance through metrics such as precision, recall, F1-score, and Area Under the Curve (AUC), This study is unique in applying ANN and RF for Card fraud detection, while integrating explanatory methods such as SHAP Our work differs from purely predictive approaches by emphasizing the importance of model interpretability; enabling the identification of the most important features that are affecting our decisions. This contribution serves as a framework for researchers and practitioners towards deploying these techniques in order to attain more robust fraud detection systems.