Few　studies　have　been　carried　out　on　the　discrimination　of　precious　Tetrastigma　hemsleyanum,　also　known　as　Sanyeqing　in　China.　Fourier　transform　near-infrared　(FT-NIR)　spectroscopy　coupled　with　chemometric　class-modelling　techniques　to　rapidly　and　effectively　discriminate　T.　hemsleyanum　was　investigated　in　this　study.　A　relevance　vector　machine　(RVM)　was　used　to　build　a　stable　and　accurate　model.　Furthermore,　competitive　adaptive　reweighted　sampling　(CARS)　was　employed　to　extract　effective　wavelength　variables.　The　results　indicated　that　the　accuracy　of　the　RVM　model　was　satisfactory　due　to　a　good　discrimination　rate.　Additionally,　the　variable　number　of　the　CARS　model　was　validly　improved　with　a　wavelength　variable　of　26,　and　the　results　of　the　RVM-CARS　model　were　also　effective.　The　results　suggested　that　the　CARS-RVM　model　is　a　suitable　model　to　rapidly　and　efficiently　discriminate　T.　hemsleyanum.　©　2017　The　Royal　Society　of　Chemistry.