Patient outcomes are greatly impacted by brain tumor detection, which is essential for early diagnosis and treatmentplanning. By evaluating magnetic resonance imaging (MRI) images and other medical imaging data, deep learning systems havedemonstrated encouraging results in automating this process. An extensive evaluation of current developments in deep learningbased techniques for brain tumor identification is provided in this work. It covers a number of topics, such as network structures,training methods, assessment metrics, and data pre-processing. In order to give researchers and doctors working towards moreprecise and effective brain tumor detection systems some insight, it also addresses issues and potential future paths in the field.The early identification and successful treatment planning of brain tumors depend on their detection. Deep learning methods, namelythose that analyses magnetic resonance imaging (MRI) scan data, have demonstrated promise in automating this process. Recentdevelopments in deep learning-based techniques for brain tumor identification are reviewed in this publication. Trainingmethodologies, network designs, evaluation metrics, and data pretreatment are just a few of the topics it covers. Convolutional neuralnetworks (CNNs), one type of deep learning algorithm, have proven to be exceptionally effective at identifying and localizing braintumors based on their ability to discern intricate patterns from magnetic resonance imaging. Prospective paths are explored andissues like data scarcity, interpretability of the model, and generalization to different populations are tackled. To further enhancethe area and advance patient care, collaboration between researchers, physicians, and industry players isKeywords :MR Images, Brain Tumor, Convolution Neural Networks (CNN), Image recognition, Deep learning, VGG-16, RESNET50.