The invention provides a method and device for outputting information, which relate to the field of image recognition, in particular to the field of intelligent transportation, and can be applied to an automatic driving scene. The method in the specific implementation scheme comprises steps of acquiring a real-time video stream, detecting key points of a running target vehicle and a parking space vehicle parked in the parking space from the real-time video stream, performing target tracking according to the key points of the target vehicle to obtain a tracking trajectory, if the tracking trajectory intersects with the area where the parking space vehicle is located, conducting coordinate conversion according to the key points of the target vehicle and the key points of the parking space vehicle, and acquiring a first area and a second area of a top view under physical coordinates, and if the first area and the second area are overlapped, outputting prompt information. According to the embodiment, the adaptability of the parking system can be improved, the functions of the on-road parking system are increased, meanwhile, traffic accidents are reduced, the traffic environment and the social environment are optimized, and development of the intelligent traffic field is actively promoted.