Sports image classification presents a significant challenge due to the wide variability in sports activities, environmental conditions, and visual similarities across different types of sports categories. This study investigates the application of Convolutional Neural Networks (CNNs) for the automated classification of sports images, leveraging a large-scale dataset containing 100 distinct sports categories. A tailored CNN architecture is developed and trained to effectively extract spatial features and perform accurate classification. To improve model generalization, a comprehensive preprocessing pipeline is implemented, including image resizing, normalization, and data augmentation techniques. The proposed models are evaluated based on training accuracy and loss metrics, demonstrating the effectiveness of deeper CNN architectures in capturing complex visual patterns. These findings underscore the potential of deep learning methodologies in supporting applications such as sports analytics, event recognition, and digital media organization.