This thesis implements a phishing detection and defense program using ensemble methods. Cybercrime is becoming increasingly sophisticated recently. The necessity of intelligent defense systems that can adapt to this growing threat is becoming more critical. In this study, I used the ensemble techniques which are VotingClassifier, BaggingClassifier, and StackingClassifier. Each ensemble technique was applied to datasets for websites, emails, and URLs. The results of the experiment showed that each method improved the accuracy of phishing detection. Finally, a user interface using Streamlit was implemented for phishing detection, and the system's defense capability against real phishing attacks was evaluated in real time.