In this paper, we study about the rear side vehicle detection system using image processing algorithm. Camera for video acquisition, controller to handle and rear-side vehicle detection algorithm was developed. In addition, the system applied to the actual vehicle and verified against the possibility of application for vehicle. Test results of this system, recognition in the stop state have good characteristics more than 95% effective recognition at the maximum 60m distance, distance error was ± 5m and relative speed error was ± 5㎞/h. Also, recognition in the drive state have 91.5%, 90.1% at day time and 72.1% at night time. Daytime recognition rate was good, but night recognition rate was relatively poor. For improvement of system performance, need algorithms improvement to raise of recognition rate, first of all, need to research for night-time recognition rate improvement.