The thesis develops a driving assistant system that can locate the positions of the lane boundaries and detects the existence of the front-vehicle. It can also provide warning mechanism so as to avoid the danger as possible collides with previous vehicle. In lane recognition, we utilized the largest gradient of luminance value to detect possible road surface marking, then cooperated with the marking static and dynamic behavior of road surface characteristic to set up road surface marking and detect system. On the other hand, we considered vehicle detection leach the vehicle bottom shade characteristic from dynamic area threshold processing, and then judge and label where the vehicle exits. By the principle of the optics image formation, we estimated the relative distance from the previous vehicle. In this thesis, we proposed an easy and fast measure for previous vehicle of 96% correct rate in different environment. Running on typical 1.7Ghz processor system results up to 80 frames per second.