This　paper　proposes　an　optimal　image　matching　model　on　the　basis　of　multi-order　features,　mainly　including　first-,　second-　and　third-order　features.　These　features　are　defined　by　a　feature　point,　an　edge　linking　two　feature　points　and　a　triangle　connecting　three　feature　points,　respectively.　The　matching　model　is　a　weighted　bipartite　graph　taking　them　as　vertices.　Its　optimal　solution,　the　maximum　weight　matching,　can　be　computed　by　the　Kuhn-Munkras　algorithm.　Experimental　results　show　that　the　model　has　good　performance　even　in　cases　of　obvious　rotation,　scale,　and　affine　transformation,　and　usually　better　than　the　Flann　and　BruteForce　algorithms.