On a global basis, brain tumors rank among the leading death factors. Because of its complexity and calm nature, this disease is challenging to diagnose in its early stages. Several types of cells give birth to brain tumors, and these cells can provide information on the type, severity, and rarity of tumors. Tumors can develop in a variety of places, and the location of a tumor may reveal information about the sort of cells that are generating it, assisting with further diagnosis. This paper suggests an innovative method to identify and segment tumor from the MRI images using Deep Learning. We used 3 pre-trained models for training our dataset of brain tumors they are ResNet50, VGG16 model 1 and VGG16 model 2. Afterward, we used Convolutional Neural Network (CNN). In our work, among all implemented models ResNet50 acquired an accuracy of 98.88%, which is really compelling. Its accuracy will help with brain tumor identification and prevention at early stages before the tumor results in any physical side effects.