The precise and timely identification of brain tumors is essential for effective treatment. However, traditional brain tumor classification methods rely heavily on human expertise, which can result in potential inaccuracies and prolonged diagnoses. To address this issue, this project aims to improve the precision and efficiency of brain tumor classification through the utilization of transfer learning techniques to refine pre-trained models using MRI scans of brain tumors. The primary goal is to establish a dependable brain tumor classification system that can aid medical professionals in promptly making precise diagnoses.