Abstract: In the current scenario most popular cyber crime today is phishing, Phishing is a social engineering-based attack where the phisher targets to retrieve the user’s legitimate confidential details to exploit and it is commonly done through sending mail or any other electronic communication in an automated fashion. In this paper a Machine Learning based approach on URLs of a website is done to detect phishing websites aiming to reduce the number of features for the model and improve the accuracy. The data-set contains numerous tested URLs which are labeled as good and bad.