- In today's world, manually examining a large number of MRI (magnetic resonance imaging)images and detecting a brain tumor is a time-consuming and incorrect task. It may have an impact on the patient's medical therapy. It might be a time-consuming task because to the large amount of image data sets involved. Because normal tissue and brain tumor cells have a lot in common in terms of appearance, segmenting tumor regions can be difficult. As a result, a highly accurate automatic tumor detection approach is required. In this study, I useda convolutional neural network to segregate brain tumors from 2D magnetic resonance brain images (MRI) and then compared the results. Moreover, I conducted the research on the six traditional classifiers namely- Support Vector Machine (SVM), K-Nearest Neighbor(KNN), Multi-layer Perceptron (MLP), Logistic Regression, Naive Bayes and Random Forest and deep learning approaches then compared with a convolutional neural network(CNN).To properly train this algorithm, I took a variety of MRI pictures with a variety of tumor sizes, locations, forms, and image intensities. We used «TensorFlow" and" Keras "in" Python" to develop the solution because it is an efficient programming language for performing rapid work. I also performed a literature review on this topic, and the study concluded with a recommendation for additional researchin this area.