Raster　binocular　stereo　vision　sensor　is　widely　used　in　3D　measurement,　and　one　of　its　difficulties　is　light-stripe　matching　and　recognizing.　Based　on　the　relationship　between　the　viewed　stripes　number　and　the　stripe　sequence　combination,　a　fast　matching　method　for　light-stripe　based　on　3D　searching　was　proposed.　The　raster　binocular　stereo　vision　system　was　regarded　as　two　raster　structured　light　sensors,　and　then　was　calibrated　under　one　common　coordinated　system.　So,　the　3D　coordinates　of　the　light　stripes　center　points　were　obtained　by　the　two　raster　structured　light　sensors,　respectively,　as　well　as　the　distances　between　the　light-stripe　center　points.　As　a　result,　the　light-stripes　whose　model　numbers　are　equal　and　have　the　minimal　distance　are　matching　stripes.　All　the　possible　combinations　were　listed　after　the　relationship　between　the　viewed　stripes　number　and　the　stripe　sequence　combination,　and　the　matching　method　was　designed　with　the　character　of　much　less　extra　calculations　which　brings　about　less　computational　burden.　The　experiment　results　show　the　effectiveness　of　the　method.