Various methods can detect lung diseases caused by viruses, and various challenges can be faced in diagnosis and treatment. However, this research aims to build a federated learning framework for lung disease detection with the help of pre-trained deep-learning models through CT images Federated learning allows multiple clients, each with its dataset, to collaboratively train a global model without exchanging specific data, thereby preserving data privacy. In this study, three local clients were established, each with its own unique dataset of CT images. These clients collaboratively develop a global model to identify lung diseases by detecting alterations in lung imaging. This automated AI-assisted software acts as a supplementary diagnostic tool. The primary objective is to significantly improve patient care by providing rapid and accurate diagnosis while minimizing the risk of disease transmission among medical personnel and the broader public. The proposed federated learning-based approach demonstrates the potential for enhanced diagnostic accuracy and improved healthcare outcomes in respiratory disease management.