Phishing attacks continue to pose a substantial and persistent threat to online security, jeopardizing the privacy and financial stability of individuals and organizations. The timely and accurate detection of these malicious websites remains an ongoing and critical challenge in the field of cybersecurity. In this research, we introduce an innovative approach aimed at addressing this issue through the utilization of Gradient Boosting Classifiers (GBCs).Phishing websites are specifically crafted to replicate the appearance of legitimate sites, with the primary intent of deceiving users into divulging sensitive information. Consequently, these malicious websites exhibit nuanced attributes that distinguish them from authentic counterparts. Traditional rule-based methods often encounter difficulties in effectively capturing these subtle distinctions. In contrast, Gradient Boosting offers an ensemble learning framework capable of harnessing the collective strength of weak classifiers, resulting in a robust model proficient at identifying these elusive features.Our experimental findings conclusively establish the superior performance of our proposed approach across various metrics, including accuracy, precision, recall, and the F1-score. Notably, our model demonstrates exceptional resilience against the adversarial tactics frequently employed to obscure the true nature of phishing websites. This research represents a significant step forward in the ongoing effort to fortify online security and protect users from the pervasive threat of phishing attacks.