Phishing emails pose a persistent security challenge, often resulting in significant financial loss and breaches of personal and corporate data security. These deceptive communications mimic legitimate entities to illicitly gather sensitive information from unsuspecting individuals, exploiting human vulnerabilities to security threats. As digital communication becomes increasingly ubiquitous, the prevalence and sophistication of phishing attacks have escalated, necessitating more advanced detection methods to safeguard information assets. This study introduces an innovative solution to the phishing quandary by employing Convolutional Neural Networks (CNN), renowned for their performance in pattern recognition tasks, in tandem with Particle Swarm Optimization (PSO), a technique inspired by social behavior in nature for optimizing computational processes. The novel integration of PSO with CNN capitalizes on PSO’s capability to fine-tune CNN hyperparameters efficiently, aiming to enhance the model’s precision in distinguishing between phishing and legitimate emails. The key findings from our experimental analysis reveal that the PSO-optimized CNN model achieves a higher accuracy and precision compared to a baseline CNN model. Notably, the optimized model demonstrated a precision of 97.27% and an ROC-AUC score of 0.991, indicating its exceptional ability to reduce false positives and accurately identify phishing attempts. The implication of these findings is profound; by significantly improving the reliability of phishing email detection, the proposed model can be instrumental in fortifying cybersecurity measures in various digital communication environments, providing a robust defense mechanism against the evolving landscape of phishing threats.