Respiratory diseases pose a significant global health challenge, necessitating accurate and efficient diagnostic tools. In this research paper, we present a novel approach to classify respiratory diseases using Convolutional Neural Networks (CNNs). We utilized a Kaggle dataset comprising 10,000 high-quality chest X-ray images distributed across five classes: bacterial pneumonia, corona virus disease, normal, tuberculosis, and viral pneumonia. Each class is composed of 2,000 images. Our methodology involves customizing the pre-trained Xception model, fine-tuning it to suit the specific requirements of respiratory disease classification. The proposed model underwent rigorous training, validation, and testing processes to ensure its effectiveness. Notably, our model achieved impressive results with an accuracy of 95.89%, precision of 95.95%, recall of 95.89%, and F1-score of 95.90%. This research showcases the potential of deep learning and CNNs in accurately classifying respiratory diseases, thereby offering a valuable tool for early diagnosis and treatment planning. The high-performance metrics attained by our model demonstrate its efficacy in the critical task of respiratory disease classification, ultimately contributing to improved healthcare outcomes.