Brain tumor refers to the unusual and uncontrolled development of brain cells within the brain. Brain tumors are rapidly overtaking all other causes of death worldwide. Effective treatment for brain tumors depends on early identification and precise diagnosis, which also improves the patient’s chance of recovery. Nowadays, deep learning models have become highly effective tools in the analysis of medical images, such as MRI scans. Hence, this work offers a comprehensive approach that incorporates survival prediction, segmentation, and classification of brain tumors. CNN is used in this study for classification. A diverse dataset of brain MRI images is employed to classify brain tumors into four distinct categories—meningioma, pituitary, glioma, and non-tumor. Additionally, the U-Net method is utilized to segment images and determine the amount of impacted regions by comparing tumor pixels to all other pixels in the image, hence yielding useful details regarding the tumor’s growth. Furthermore, a CNN-based approach provides survival prediction. The accuracies achieved for classification, segmentation, and survival prediction based on this approach are 98.12%, 92%, 96% respectively. This integrated strategy provides a comprehensive solution to improve the efficiency of brain tumor management by combining classification, segmentation, and survival prediction.