Stable　video　has　a　significant　effect　on　video　based　traffic　parameters　extraction　system.　This　paper　proposed　a　fast　video　stabilization　method　based　on　feature　matching　and　histogram　clustering.　Firstly,　the　method　extract　ORB　feature　points　on　sub-images　which　are　set　in　adjacent　frames　and　then　matches　them　based　on　Hamming　Distance.　Secondly,　the　motion　consistency　principle　combined　with　histogram　clustering　is　used　to　get　correct　matches　which　can　be　used　to　calculate　the　parameters　of　the　affine　transformation　matrix.　Finally,　after　filtering　those　parameters,　stable　frame　can　be　obtained　by　warping　the　original　frame　using　the　affine　transformation　matrix.　Experimental　results　show　that　the　proposed　method　can　effectively　stabilize　videos　with　low　complexity　in　real　time.