Traditional　Spectral　detection　technology　has　shortcomings　such　as　strong　subjectivity,　time-consuming,　complicated　operation,　etc.　In　this　paper,　we　constructed　an　Anoectochilus　strains　classification　model　based　on　Adaboost　ensemble　learning　by　LinkSquare　®　NIR　hand-held　spectrometer.　The　experimental　research　results　show　that　the　accuracy　rate　of　Adaboost　ensemble　learning　can　still　reach　95.6%　under　the　influence　of　the　detection　environment　and　experimental　conditions　of　the　NIR　hand-held　spectrometer.　This　method　can　effectively　solve　the　problems　of　the　small　discriminability　of　the　leaf　morphology　of　Anoectochilus　roxburghii　and　the　low　accuracy　of　a　single　classification　model.　It　provides　a　convenient,　fast　and　effective　detection　method　for　the　detection　of　Anoectochilus　roxburghii　strains.　©　2022,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.