As the number of cars continues to increase, parking has become a major challenge in traffic management. The act of drivers searching for parking spaces in urban areas not only wastes time but also exacerbates traffic congestion and increases carbon emissions. Therefore, this study proposes a parking management system based on deep learning technology. The system collects training data through drones and utilizes YOLOv9 to enhance recognition accuracy. Experimental results demonstrate that the model achieved high accuracy in vehicle recognition, with a precision of 99.74%, recall of 99.84%, and an F1 score of 99.78%. Finally, a web interface provides navigation services for users, allowing them to easily locate the nearest available parking space and navigate directly to it, effectively reducing search time and environmental impact.