Brain tumors and multiple sclerosis (MS) are significant medical conditions with overlapping clinical features, posing diagnostic challenges. Brain tumors, whether benign or malignant, compress the brain’s surrounding tissues and can cause a variety of neurological conditions. Traditional treatments include surgery, medication, and radiation, with a survival rate of 35% for malignant cases in the U.S. About one million people worldwide suffer from multiple sclerosis, a chronic CNS disease that causes intermittent neurological symptoms and long-term disability. Recent advancements in MS treatment focus on immune system modulation with new oral and infused therapies. Due to its tumorlike appearances, which make it difficult to distinguish MS from brain tumors, diagnosis can be especially difficult in this case. This research provides a novel method to classify brain MRI scans into three categories: multiple sclerosis, brain tumor, and normal using a Convolutional Neural Network (CNN) linked with an attention mechanism. Achieving an accuracy of 96%, this method addresses the diagnostic difficulties posed by the similarity in imaging features between tumors and MS. The proposed classification model highlights the novelty of integrating these specific classifications and provides a robust solution for distinguishing between these complex conditions.