A brain tumor is a cluster or group of anomalous cells in the brain that can become life-threatening due to its ability to dominate neighboring tissues along with forming metastases. An exact diagnosis is necessary for treatment planning to be successful. However, existing classification methods often struggle with inaccuracies and early detection due to variations in tumor shape, size, and intensity across different magnetic resonance imaging (MRI) scans. Hence, this research proposes Deep Belief-Visual Geometry Group-16 (DB-VGG-16) for brain tumor classification utilizing MRI. Initially, the MRI image is acquired from the dataset, and image preprocessing is effectuated with the help of a bilateral filter to denoise the image. After the filtered image is fed as input to the image segmentation process, the tumor-affected region is segmented by applying morphological operations. Later, features are mined from the tumor-affected region, and this is established using statistical and texture features. Ultimately, the brain tumor classification is achieved by utilizing the designed DB-VGG-16, which is devised using deep belief network (DBN) and Visual Geometry Group-16 (VGG-16). The evaluation is conducted using multimodal brain tumor segmentation challenge 2018 (BraTS) and figshare datasets using metrics, such as accuracy, specificity, sensitivity, F-measure, and precision. The DB-VGG-16 recorded a maximal specificity of 0.918, accuracy of 0.928, sensitivity of 0.903, precision of 0.916, and F1-score of 0.910 using figshare dataset with K-Fold-9.