Phishing attacks on websites are a type of cyberattack in which attackers create fraudulent websites that imitate legitimate platforms such as social media in order to deceive unsuspecting users and disclose confidential information. This includes passwords, credit card data, user ID, and other personal data. These phishing websites look real and often use various methods such as URL swapping, social engineering, and email or text phishing to attract victims to reveal their sensitive information. Web applications are becoming more and more complex and difficult to detect at first glance, especially when using encryption and obfuscation methods. Machine learning needs to be developed to effectively detect and prevent phishing web applications from being uploaded to the server in real time. In addition to analyzing machine learning algorithms to detect attacks based on web applications, the study calibrates new analyses by executing machine learning algorithms and validating the results. The study uses unique and categorized results from machine learning data sets. According to the results obtained from the experimental and comparative analysis of the classification algorithms used, the random forest model showed the highest accuracy, with an impressive 97%, followed by the decision tree model of 94% and extreme gradient amplification (XGBoost).