<p>In the world of emails, spam messages present a significant challenge, leading to inconveniences and potential security risks. Addressing this issue, the task of spotting spam in emails is critical for ensuring secure and trustworthy communication. However, two prevalent approaches have their challenges. The centralized model, gathering data in one place, raises privacy issues. Conversely, the federated learning model, which focuses on privacy, can lead to a compromise in accuracy. This research paper presents a novel federated learning-based fair clustering technique for spam email detection. By addressing privacy concerns and aiming for accurate classification, the proposed approach Fair Clustering model combines the strengths of federated learning and data clustering. Through experimental evaluation, the Fair Clustering model is evaluated against both a centralized and federated learning model. Different metrics, such as accuracy, recall, precision, and F1-score, are used to evaluate and compare the performance of these models. The results demonstrate that the Fair Clustering model outperforms the federated learning model, showcasing the effectiveness of fair clustering in selecting representative clients and improving classification performance.</p>