A brain tumor is abnormal cells that develop in the brain structure. Brain tumors grow can create pressure and affect the functions of surrounding brain tissues, which is life-threatening. In this work, we propose an approach using deep learning networks to detect and segment brain tumors with 3D U-Net by changing the backbone with MobileNet-V2, ResNet-101, and DensetNet-121. We used the BraTS 2020 dataset for conducting experiments. The results show that the proposed method achieves high accuracy of up to 99%. We use the 3D Slicer tool in modeling 3D brain tumors to effectively support doctors in the early detection and classification of brain tumors for timely treatment.