The demand for additional parking spaces in current urban settings is a significant issue in modern parking management. As cities grow, the optimization of parking resource utilization becomes increasingly challenging, which in turn leads to more traffic and pollution. This paper presents a real-time Smart Parking System that will help identify the occupancy of parking spaces through deep learning methods. YOLO object detection is applied in implementing this system to recognize parking spaces as either occupied or empty. It leverages the remarkable speed and accuracy of YOLOv8 to process video feeds from parking lots in real-time. This approach involves training the YOLOv8 model on large data, which consists of images and real-time videos of parking lots. The outcome of the YOLOv8 model is measured with precision and recall. Results show that the model achieved a remarkable 99.9% precision on available datasets and 98.9% precision on self-prepared datasets.