This paper introduces an approach to harnessing the power of Artificial Intelligence (AI) to enhance the sustainability of healthcare systems in smart cities. The convergence of AI and healthcare holds immense potential to address existing challenges. This study explores the unification of Internet-of-Medical-Things (IoMT) framework and Federated Learning (FL) the to foster sustainable healthcare. The focus is on privacy-preserving healthcare data analysis by leveraging FL to process data at its source, thus reducing communication costs while minimizing the risk of transfer of sensitive data. This approach is applied within the IoMT ecosystem, optimizing and processing COVID-related CTS through FL. The process involves training local models, which are then aggregated using Federated Averaging (FedAvg) to create a comprehensive global model. The proposed approach ensures model accuracy over time and upholds data privacy, which is crucial for sustainable healthcare. Results show that with 10–20 clients, the model achieves a global accuracy between 94.5% and 94.7%, demonstrating robustness and scalability.