The persistent threat of financial loss, unauthorized access to sensitive information, and identity theft through phishing attempts using malicious URLs in emails necessitates continual advancements in detection methods. This research addresses this challenge by proposing and evaluating innovative machine learning and deep learning algorithms—namely, Decision Tree, Random Forest, Multilayer Perceptron, and XGBoost—aimed at fortifying our defenses against the expanding landscape of phishing threats. This paper centers on augmenting the efficacy of existing algorithms through advanced feature extraction methods and a meticulous refinement of feature collection. In a comparative study, XGBoost emerged as the top performer across key metrics, including accuracy, precision, recall, F1 score, and Negative Predictive Value (NPV). Conversely, Random Forest exhibited commendable performance in minimizing False Discovery Rate (FDR) and False Positive Rate (FPR). Beyond benchmarking these algorithms, our research strives to contribute to the cybersecurity domain by shedding light on their practical application in email security systems. By comprehensively assessing the strengths and weaknesses of machine learning and deep learning models in identifying phishing URLs, our work aims to significantly bolster internet safety for users.