Tuberculosis (TB) is still a leading global infectious disease, therefore early and accurate diagnosis is critical for the effective treatment and control of the disease. The paper presents a deep learning based approach for automatic classification of chest X-ray images into Normal and Tuberculosis (TB) classes. The experiments were performed on a publicly available Kaggle Tuberculosis Chest X-ray Dataset consisting of 3,500 Normal images and 700 TB images. We developed a custom Convolutional Neural Network (CNN) to learn discriminative features from chest X-ray images automatically. The preprocessing pipeline consisted of resizing images to 224 × 224 pixels, normalizing the pixels, and noise reduction. We applied data augmentation techniques to help the model generalize better and to mitigate the problem of class imbalance. The CNN was trained with Adam optimizer, binary cross entropy as loss function, learning rate scheduling and early stopping Experimental results demonstrate that the proposed CNN achieved a classification accuracy of 96.07% on the test dataset. The findings demonstrate the potential of the proposed CNN-based approach for automated TB screening from chest X-ray images and provide a foundation for developing computer-aided diagnostic systems, particularly in resource-constrained healthcare environments.