For medical analysis and interpretation, MR brain image categorization that is automated and accurate is essential. The most significant part of the human body that MRI, or magnetic resonance imaging, allows us to see in great detail is the brain. It provides very important information that helps in the clinical diagnosis of tumors and helps in biomedical research field. Tumors can be detected through ultrasound techniques, CT scans and MRIs. However, using the MRI images would prove to be fruitful because the accuracy rate will increase. AI can help in providing precise information about brain tumor to doctors. In this paper, we presented a brain tumor detection system on pre-deep learning and during deep learning era. From the study, it is clear evident that compared to pre-deep learning, deep learning era produces a better results. The performance of the suggested models are evaluated in terms of sensitivity, accuracy, efficiency, and specificity.