For　head　pose　estimation　plays　an　important　role　in　human　machine　interface,　the　authors　propose　an　approach　to　estimate　3D　head　pose　from　a　monocular　sequence.　A　block　matching　criterion　is　designed　and　the　effect　of　block　matching　for　feature　tracking　is　improved.　An　algorithm　based　on　facial　geometry　information　and　projective　model　is　presented　to　estimate　the　head　pose.　The　experimental　results　show　that　our　method　is　effective　for　head　pose　estimation.