Over time, fraudulent online transactions have resulted in significant losses and harm for both individuals and businesses. The proliferation of advanced technology and global connectivity has contributed to an increase in the number of cases of online fraud. Strong fraud detection systems must be created in order to counteract these losses. The correct identification of fraudulent transactions depends heavily on machine learning methods. Nevertheless, there are obstacles to overcome when putting fraud detection algorithms into practice, including uneven class distributions, sensitive data, and restricted data availability. The secrecy of records makes it more difficult to make deductions and create better models in this field. By utilizing the Credit Card Fraud dataset, this study investigates various algorithms such as Logistic Regression, Random Forest, XGBoost that are appropriate for categorizing transactions as either authentic or fraudulent. By comparing all three algorithms XGBoost gives better accuracy of 98%.