Tuberculosis is a chronic lung illness. The bacteria spread through the air and are the cause of widespread illness. Early detection can prevent it from spreading to other organs such as the kidney, spine, and brain, lowering the likelihood of the disease being harmful. However, human error could occur during radiologists’ manual diagnosis of patients utilizing chest X-rays Because of this, researchers have invested a lot of effort into developing an automated decision support system that can accurately identify tuberculosis from chest X-ray images. To address this problem and save many lives, a convolutional neural network-based deep learning model can detect tuberculosis. Establishing computer-aided diagnostics (CAD) for tuberculosis early diagnosis. Deep learning is being applied to get beyond the constraints of clinical diagnosis. Many DL techniques in use today work incredibly well, and in the instance of image analysis, CNN (Convolutional Neural Network) is stronger for feature extraction and producing predictions. It is the best way of increasing production in any sector. A CNN is an illustration of a neural network and is widely utilized in vision in computer research. Four different Pre-trained models are used DenseNet-169, MobileNet, Xception, and Inception-V. DenseNet performs well compared to other models for the selected Dataset.