One of the most challenging tasks in the analysis of medical images is to identify brain tumors correctly. Due to its intricate nature of diagnosis method and the wide range of tumor sensory tissues, the process is still tedious. Hence, it is necessary to enhance various brain tumor identification methods for medical applications. A vital phase in analysing medical imaging is segmenting a brain tumor. The possibility of effective therapy and the likelihood that patients will survive are both significantly improved by early identification of brain tumors. It is really tedious and more time taken to segment huge amount of MRI data for the analysis of abnormal brain acquired during routine clinical procedures. Therefore, automatic brain tumor segmentation is demanding these days. This paper present an systematic survey of segmentation as well as categorization methodologies based on MRI scans. Methods based on deep learning have currently gained popularity in the field of segmentation and classification because they generate innovative outcomes and are preferable to prior techniques in tackling the issue. These techniques enable efficient processing and unbiased evaluation of the various MRI data samples. This study highlights future research developments for integrating segmentation and classification methods of brain scans into standard clinical practice.