In MRI-based healthcare systems, precise brain tumor classification is essential for assisting with medical evaluations and treatment planning. We present an effective approach for classifying brain tumors into four target classes: glioma, meningioma, pituitary tumor, and no tumor. By utilizing a convolutional neural network (CNN) to extract features from MRI scans and combining these features with conventional machine learning classifiers like Random Forest, K-Nearest Neighbor, and Support Vector Machine, this method achieves a notable improvement in tumor detection accuracy. Furthermore, to optimize classification performance, a stacking classifier ensemble method is utilized, capitalizing on the strengths of the individual models. Experimental results demonstrate the proposed method’s efficacy, achieving an accuracy of 97% and providing a reliable solution for precise brain tumor identification in clinical applications.