Fraud refers to criminal deception that convinces victims to reveal personal information such as their password or credit card number. Fraudulent websites are usually designed to appear professional and convincing, as if they are genuine. To mitigate the negative effects of a fraudulent website. We proposed an effective phishing website detection system that analyses phishing website URL addresses in order to learn data patterns that can distinguish between authentic and phishing websites. To learn data patterns in website URLs, our system uses machine learning techniques such as decision trees. Using a random forest classifier, we evaluate our system on a recent phishing website dataset.