Brain tumours constitute 8 to 12% of all juvenile malignancies in India, compared to roughly 21% in the West. A brain tumour occurs when abnormally grown cells in the brain or adjacent organs, such as nerves, pituitary gland, or pineal gland, form masses. Diagnosing brain tumours typically involves MRI scans to detect these abnormal masses. Detecting brain tumours usually starts with Magnetic Resonance Imaging (MRI) for detailed images and Positron Emission Tomography to check tumour activity. Traditional methods for detecting brain tumours involve manual segmentation and feature extraction from MRI images, which can be prone to errors. In response, researchers have taken a unique approach, developing a Convolutional Neural Network (CNN) paired with a Decision Tree for brain tumour diagnosis. This hybrid model outperformed the other methods regarding predictive value, sensitivity, and classification accuracy. The improvement is due to incorporating a Decision Tree (DT) rather than the typical fully connected layers often observed in Convolutional Neural Networks (CNNs). The Novelty of this method, evaluated with MRI images, has achieved a remarkable accuracy rate of 98%, outperforming the existing Multiclass CNN-SVM technique. This advanced methodology offers significant potential in aiding physicians with early tumour detection and guiding critical treatment decisions, thereby potentially reducing patient mortality rates and improving overall healthcare outcomes, providing substantial benefits to the medical community.