Brain tumor diagnosis and treatment create critical challenges due to the complexity of their nature. Deep learning techniques have appeared as reliable tools for the accurate identification and classification of brain tumors from MR images. This paper presents a novel approach utilizing the pretrained deep learning Xception model, to classify MRI images of brain tumors. The proposed method achieves an accuracy of 99%, which is far better than existing methods. Through comprehensive evaluation and experimentation, the research presents the effectiveness of the proposed CNN model. Our proposed approach can serve as an example for conducting and modeling other research. Furthermore, the paper explains the potential implications of deep learning in medical imaging and highlights avenues for future research.