Reducing the risk pose by phishers and other cyber criminals in the cyber space requires a robust and automatic means of detecting phishing websites, since the culprits are constantly coming up with new techniques of achieving their goals almost on daily basis. Phishers are constantly evolving the methods they used for luring user to revealing their sensitive information. Many methods have been proposed in past for phishing detection. But the quest for better solution is still on. This research covers the development of phishing website model based on different algorithms with different set of features. The evaluation criteria are used in measuring the performance of phishing detection. Benchmark phishing website dataset were considered in the experiment. The result of the experiments showed that XGBOOST is better in most of the problems than the other methods in terms of the F.score, MCC, and Accuracy Therefore, the proposed method represents a very competitive for phishing detection. XGBOOST has a better regularization ability which helps to reduce overfitting, high speed and flexibility due to it costume optimization objectives and evaluation criteria.