SARS CoV, widely known as COVID-19, is a pandemic that affected the whole world. It causes respiratory problems, and illness, especially for those with other co-morbidities. Limited resources for detection and testing kits during the pandemic lead to serious illness. The symptoms are challenging to identify leading to severe complications. Artificial models help in diagnosing various medical diseases at a great rate. Deep learning models are more efficient in healthcare sectors. The proposed method involves CNN-DenseNet for classification of COVID-19. This model uses ResNet, XceptionNet, and DenseNet to diagnose COVID using chest X-ray images. This model achieves an accuracy of 98.29% on classification on hybrid CNN-DenseNet. The accuracy and loss of the model are calculated. With all the models, CNN-DenseNet provides higher accuracy than traditional models.