Detection of brain tumors in medical imaging is essential \or the early diagnosis and treatment planning. Image classification applications, including medical image analysis, have demonstrated the effectiveness of conventional convolutional neural networks (CNNs). Nevertheless, they encounter challenges in identifying objects with significant fluctuations in size, shape, or location, which are common in medical images, such as brain tumor imaging. In recent times, deep learning models have demonstrated encouraging outcomes in several medical imaging applications, including brain tumor detection. The capacity of capsule networks to capture spatial relationships while accounting for changes in object placement and scale has piqued interest. These networks are a novel type of neural network architecture. In this scenario, we suggest the utilization of capsule networks and deep learning to diagnose brain lesions. Our approach commences with the preprocessing of brain MRI scans to extract pertinent information, which is subsequently followed by the training of a capsule network to identify tumors and non-tumors in the images. The capsule network’s goal is to capture the hierarchical structure of objects in photographs, thereby enabling the network to generate meaningful representations of brain tumor features. Furthermore, we employ data augmentation techniques to enhance the model’s generalization performance and resilience.