Pharming is a two-pronged cyberthreat that involves installing harmful malware on a victim’s device and then redirecting them to false websites. It is a serious threat to internet security. This work investigates how to improve the detection of pharming assaults by applying machine learning (ML) techniques, specifically focusing on Naive Bayes classification algorithms. By using principal component analysis (PCA) to classify URLs according to their features, the suggested model offers a thorough summary of the dataset that was acquired from UCI and consists of typical pharming attack cases. Using machine learning (ML) to analyze large datasets and find traits, trends, and anomalies that point to malicious activity is the methodology. The suggested method is made more effective by exploring the dataset’s attributes through the use of feature.