Phishing attacks pose a significant threat to internet users by attempting to deceive them into divulging sensitive information such as passwords, credit card numbers, or personal data. Traditional phishing detection methods often rely on static blacklists or heuristic rules, which may not effectively capture the evolving tactics employed by malicious actors. Machine learning (ML) techniques offer a promising approach to enhance phishing detection by learning patterns from large datasets. This paper presents a comprehensive overview of utilizing machine learning algorithms for detecting phishing websites, from data preprocessing to model evaluation. We explore various features, algorithms, and evaluation metrics commonly used in the field. Additionally, we discuss challenges, future directions, and potential countermeasures to improve the effectiveness of phishing website detection using machine learning.