The invention discloses an intelligent prediction system for automobile driving. The intelligent prediction system comprises a traffic information processing module, a path planning module, a cloud computing module and a feedback control instruction-output module. The intelligent prediction system is characterized in that the traffic information processing module comprises a GPS positioning system, a millimeter wave radar camera, a 4G network and environment perception, wherein the environment perception comprises pedestrian recognition, vehicle recognition, obstacle recognition and target fusion. Information is sent to the cloud computing module in real time through the traffic information processing module and calculated by the cloud computing module, and the information is processed andsaved. The information is sent to the path planning module through the cloud computing module, then the feedback control instruction-output module is operated, the automobile driving can be intelligently predicted, and by using the pedestrian recognition, the vehicle recognition, the obstacle recognition and the target fusion, the driving speed and track of the automobile can be changed, so thatit is ensured that the automobile can predict unknown accidents during normal driving, and the safety of passengers and the driver in the automobile is ensured.