To reduce the accident, technology of computer vision and the iPhone are used to develop of forward collision warning system. Generally, hardware device was used for the system, such as infrared. Technology of computer vision was used for lane, vehicle and distance. Lane data was detected from features about line. Lane data was used for average method, make lane detection robust. Based on the shadow of vehicle, the vehicle position was detected. Warning machine is judges dangerous for vehicle. The warning mechanism depends on the condition that is sorted. Each country has prescribed a length of dashed line in highway. Based on length of dashed line, 4 meters of the area in front of the vehicle is calculated based on Pinhole camera model. This system can be applied to iPhone. The iPhone has a best camera capabilities, auto-exposure, and auto-correction. Experimental results show that rate of lane detection does not affect the rate of error alert in warning mechanism. Forward collision warning system is robust.