Abstarct: In recent years a growing body research has been devoted to the issue of vehicle safety. Road detection systems are among new facilities improving a driver safety and alerting the driver when a vehicle strays it however the challenges such as lane appearance diversity, variation in clarity of image, changes in visibility conditions are still problematic for roadways detection systems. In the present project, a new and efficient architecture is presented for road boundary detection. Opposed to previous drawings, in the current research the Sobel edge detection was used which improves edge detection and noise reduction by finding strong edges. After that Hough transform and Connected Component analysis were used for finding the lines. In the first phase, circuit was implemented in the Matlab simulation environment and the operation of the proposed algorithm was evaluated. After the success of the initial design, System Generator graphic simulator and HDL Coder simulator were used to simulate the design on FPGA. The simulation results on FPGA, the average execution speed is 20ms and the detection rate is 87% were obtained the proposed system. According to these results can be said to have achieved a detection rate acceptable short running time.