Lung disease is currently increasingly prevalent and has a significant impact on death rates across the world. Lung diseases are caused by airway diseases, lung tissue diseases, and lung circulation diseases. Chest X-ray-based diagnosis system is commonly used for the primary detection of lung diseases. X-ray image-based solutions are used to find the cause and severity of disease. Machine learning and deep-learning-based different image processing techniques are used to detect and classify different lung diseases. The accuracy of automatic disease detection depends on the number of samples and features of a dataset. In this work, we have used medium-volume datasets of various diseases extracted from the Kaggle repository. We address the issue of lung disease diagnosis from chest X-ray pictures for the sustainable health sector. This work represents the classification of pneumonia, tuberculosis, and COVID-19 lung diseases using a hybrid model. Concerning the dataset for diseases, we implemented three models—Convolutional Neural Network (CNN), ResNet-50, and Inception-V3. Among these CNN hybrid model outperformed the others with accuracy and recall values of 0.8766 and 0.9949 on the test set. We select the disease dataset, combine the training and validation data from that dataset, and divide all images (data) into an 80:20 ratio for training data and validation data to obtain more images for training data in the hybrid model, we constructed three CNN models that simultaneously effectively identify pneumonia, tuberculosis, and COVID-19. The hybrid model had the highest accuracy and precision for tuberculosis disease, with 0.9583 and 0.9508, respectively.