An automated system using deep learning and techniques employing three-dimensional (3D) volumetric images, such as MRI, can effectively detect and classify brain tumor MRI images, including T1 weighted, T2 weighted, diffusion, and others, which are preprocessed through bilateral filtering, normalization, or rescaling before being classified by a convolutional neural network (CNN). The model is trained on a dataset with four tumor classes: meningioma, glioma, pituitary tumor, and no tumor. Data augmentation is used for generalization, and the evaluation metrics include accuracy, sensitivity, and specificity. Besides, a graphical user interface was developed to allow users to upload images for real-time prediction. As a result, given its significant degree of accuracy, the system could assist doctors in diagnosing brain tumors with high reliability and efficiency. Our aim was to develop a deep learning technique to classify brain tumor MRI images using a CNN. The proposed CNN algorithm was applied for multiclass classification of brain tumor MRI images. The proposed model tested and compared with extensively used models such as KNNs. The results show that the hybrid approach achieves an accuracy of up to 96%, thus proving its superiority to the other models. In conclusion, deep learning models are trained with MRI images of brain tumors. There is a wide scope of implementing such models in the medical domain to aid in the detection and diagnosis of brain tumors.