This paper presents an ensemble machine learning approach for enhancing credit card fraud detection. The study addresses the challenges of data imbalance, false positives/negatives, and real-time processing in credit card fraud detection. The proposed ensemble model integrates Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Random Forest (RF), Bagging, and Boosting classifiers. To tackle data imbalance, under-sampling and the Synthetic Minority Over-sampling Technique (SMOTE) are applied. The model is evaluated using a dataset of European credit card transactions, providing a realistic scenario for assessment. The methodology includes data pre-processing, feature engineering, model selection, and evaluation. Google Colab is used for efficient model training and testing. Comparative analysis shows that the ensemble model outperforms traditional machine learning methods and individual classifiers in terms of accuracy, precision, recall, and F1-score. The study highlights the effectiveness of ensemble methods in mitigating credit card fraud detection challenges. The findings contribute to the development of more resilient and adaptive fraud detection systems. Keywords: credit card fraud detection; ensemble model; machine learning; data imbalance; Synthetic Minority Over-sampling Technique; deep learning.This paper presents an ensemble machine learning approach for enhancing credit card fraud detection. The study addresses the challenges of data imbalance, false positives/negatives, and real-time processing in credit card fraud detection. The proposed ensemble model integrates Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Random Forest (RF), Bagging, and Boosting classifiers. To tackle data imbalance, under-sampling and the Synthetic Minority Over-sampling Technique (SMOTE) are applied. The model is evaluated using a dataset of European credit card transactions, providing a realistic scenario for assessment. The methodology includes data pre-processing, feature engineering, model selection, and evaluation. Google Colab is used for efficient model training and testing. Comparative analysis shows that the ensemble model outperforms traditional machine learning methods and individual classifiers in terms of accuracy, precision, recall, and F1-score. The study highlights the effectiveness of ensemble methods in mitigating credit card fraud detection challenges. The findings contribute to the development of more resilient and adaptive fraud detection systems. Keywords: credit card fraud detection; ensemble model; machine learning; data imbalance; Synthetic Minority Over-sampling Technique; deep learning.