Drows inessorfatigueisamajorcause of road accidents. A system to detect drows iness has been developed using an approach based on vision. This approach is non-intrusive. The system makes use of a small camera that captures the driver’sface and monitors the eyes and mouth of the driver to detectd row siness. The driver’s fatigue will be detected based on blink-ing, yawning and eye closure rate.It uses video input from thecamera. Further, we used an approach to detect regions on face based on 68 facial land marks points.Once the region of interestis located, compute the change indistances between the land mark point. This determines if the eyes and mouth are open or closed. If the eyes are found to be closed for 12 consecutive frames,it considers the driver to be drowsy. Then provides a warningsound.The system operates successfully with 92% accuracy.