In the last few years, utilizing deep learning techniques for predicting tumor presence on brain MR images became quite common. By using deep learning networks, the detection of brain tumors with different segmentation techniques can be applied. In this paper, Brain Tumor Segmentation (BraTS) image data set was fed into a convolutional neural network as an input and then pixel classification operation was executed on MR images by using semantic segmentation with the network. In addition, two different classes including ‘background’ and ‘tumor’ were indicated to classify pixels. In order to evaluate semantic segmentation results, some metrics such as mean IoU and mean BFscore were used. On the other hand, ground truth labels in original image data and predicted labels which were acquired at the last layer of our network were compared by using dice score. As a final stage, 3D imaging of both brain and tumor with ground truth and predicted labels was carried out. In conclusion, by applying semantic segmentation on a test image, tumor prediction with high accuracy was obtained. It can be said that 3D imaging of brain tumors can provide surgeons visual contribution in terms of better planning regarding width, height, and depth of the brain tumor. At the same time, calculation of tumor volume can offer a significant numerical advantage in terms of tumor size, region, and treatment in addition to visual analysis.