Effective communication via email is indispensable in modern professional and per sonal contexts, facilitating efficient information exchange, collaboration, and connec tivity across global networks. Phishing emails pose a significant threat by potentially compromising personal and financial information, leading to identity theft, financial loss, and erosion of trust in online communication channels. Phishing emails pose a persistent threat to cybersecurity, necessitating robust detection, and classification mechanisms. This study evaluates the effectiveness of RandomForest and Naive Bayes classifiers in identifying phishing emails, employing the Natural Language Processing (NLP) techniques namely Term Frequency-Inverse Document Frequency (TF-IDF) and CountVectorizer methods for feature extraction, respectively. The RandomForest classifier demonstrates superior performance, achieving an impres sive accuracy rate of 98.91% with a precision score of 98.09%. In contrast, the Naive Bayes classifier yielded slightly lower accuracy at 98.44% and a precision score of 98.06%. These findings underscore the critical role of feature extraction techniques in enhancing the accuracy of phishing email detection. By leveraging machine learning algorithms and advanced feature extraction methods of NLP, organizations can bolster their defenses against phishing attacks, thereby safeguarding sensitive infor mation and mitigating potential risks.