Many financial payment systems have to face fraudulent activities due to the fast-paced development of the technology. Fraud detection is essential for the proper management of fraud control. It automates the manual checking processes and helps the detection be done conveniently. It is important to research and find ways and means of proper methodologies which will help serve the purpose of fraud detection effectively. Machine Learning Approach becomes more popular and accurate compared to a rule-based approach in this scenario. This paper presents such a performance comparison among a few methods which were tested with a dataset.