Face　recognition　with　assistance　of　3D　models　has　been　a　successful　approach　recently.　In　this　paper,　we　develop　a　face　recognition　system　fusing　2D　and　3D　face　information.　First,　the　HarrLBP　representation　is　proposed　to　represent　the　2D　faces.　Then,　the　3D　morphable　model　(3DMM)　is　employed　to　estimate　the　3D　shape　for　the　given　2D　face,　and　five　kinds　of　3D　facial　geometrical　features　are　extracted　from　the　virtual　3D　facial　meshes　to　assist　the　face　recognition.　Finally,　we　fuse　the　2D　HarrLBP　and　the　five　3D　features　under　a　linear　self-adaptive　weight　scheme　to　promote　the　final　recognition　efficiency.　The　experimental　results　on　ORL　and　JAFFE2　face　database　show　the　good　performance　of　the　proposed　fusion　method,　and　demonstrate　that　our　method　is　robust　to　the　facial　expressions　and　poses　to　a　certain　extent.