In　recent　years,　the　number　of　car　has　been　increasing,　and　car　has　gradually　become　an　indispensable　mean　of　transportation　for　people　to　travel.　However,　along　with　the　rapid　growth　of　car　brings　convenience　to　people＇s　life,　it　also　brings　many　negative　effects　to　the　development　of　road　traffic.　More　and　more　traffic　accidents　have　happened,　and　most　traffic　accidents　are　caused　by　fatigue　driving.　In　order　to　reduce　traffic　accidents　caused　by　fatigue　driving,　many　methods　have　been　proposed.　However,　these　methods　cannot　guarantee　the　accuracy　and　speed　of　detection　at　the　same　time.　So,　a　fatigue　driving　detection　method　based　on　multi　feature　fusion　is　presented　in　this　paper.　Firstly,　MTCNN　is　used　to　improve　the　face　tracking　algorithm　based　on　MedianFlow.　Then　a　new　face　key　points　detection　model　based　on　CNN　is　proposed,　the　result　of　face　key　points　detection　can　be　used　to　locate　the　eyes.　Finally,　information　such　as　eye　closing　time,　blinking　frequency　and　head　position　are　fused　to　detect　fatigue　driving.　Experimental　results　show　that　the　fatigue　driving　detection　method　proposed　in　this　paper　has　a　good　result　on　speed　and　accuracy.　©　2019　IEEE.