In this study, we present the use of lexical characteristics, blacklists, DNS data, and web-based properties of domain names to identify dangerous domains. We will employ three well-known machine learning ensemble classifiers—Random Forest, Lite GBM, and XGBoost—using the attributes retrieved from the domain names to distinguish between benign and malicious domains. Active DNS analysis is the foundation of our experiment. It is now feasible to foresee such websites, nevertheless, by employing machine learning algorithms on vast datasets. It is possible to identify harmful websites and alert visitors to the risk before they visit them using classifiers developed using methods like logistic regression and unsupervised machine learning. Using the Kaggle Malicious and Benign Website Dataset, this paper focuses on applying a range of unusual classification techniques to determine if a website is harmful or not.