Detection fraud is a crucial task which demands robust ML models for detection of fraud during transactions with more precision. The paper presents a comparative analytical view of some ML algorithms using the 2013 European cardholders dataset which is highly imbalanced. In the performance evaluation, this study considers six models based on diverse performance measures (Carrington et al. in IEEE Trans Pattern Anal Mach Intell 45:329–341, 2023). Thus, six ML models are adopted (Nandipati, Boddala in Credit Card Approval Prediction: A comparative analysis between Logistic Regression, KNN, Decision Trees, Random Forest, XGBoost 2024). An imbalanced nature of dataset introduced above requires a special assessment of the model’s capacity in terms of handling a skewed distribution of fraudulent versus non-fraudulent transactions (Mishra, Ghorpade in Credit card fraud detection on the skewed data using various classification and ensemble techniques, pp. 1–5, 2018). As per our results, we conclude that highest accuracy is achieved for models such as XGBoost and Random Forest, making them the best at identifying credit card fraud in this given dataset. The major outcome from this paper offers insight into a given machine learning technique’s ability and inadequacies as regards its ability to detect fraud and provide guidance to practitioners in finance.