In the medical imaging field, brain tumor detection is a very critical process which requires accurate localization technique in the early stage of detection in the treatment process cycle. This work presents an early diagnosis of brain tumor from MRI images by implementing computer vision technique called YOLO (you only look once). Here, YOLO is an exceedingly sophisticated AI technique which is best known for its accuracy is implemented on brain tumor image dataset in its optimizing stage by fine tuning the hyperparameters like learning rate, batch size and epochs etc. The metrics like precision, recall and mean Average Precision (mAP) model performance decides the model performance. The model achieves the high accuracy by achieving mAP of 0.771 at 50% IoU and 0.489 across multiple IoU thresholds (mAP50-95). The outcomes present the YOLOv8 capability as a diagnostic tool in the field of radiology in the tumor detection process. Future work will include the model extension to tumor segmentation by using boundary information.