Near-infrared spectroscopy (NIRS) is a potential, field-portable wood identification tool. NIRS has been studied as tool to identify some woods but has not been tested for Dalbergia. This study explored the efficacy of hand-held NIRS technology to discriminate, using multivariate analysis, the spectra of some high-value Dalbergia wood species: D. decipularis, D. sissoo, D. stevensonii, D. latifolia, D. retusa, all of which are listed in CITES Appendix II, and D. nigra, which is listed in CITES Appendix I. Identification models developed using partial least squares discriminant analysis (PLS-DA) and soft independent modeling by class analogy (SIMCA) were compared regarding their ability to answer two sets of identification questions. The first is the identification of each Dalbergia species among the group of the six above, and the second is the separation of D. nigra from a single group comprising the other species, grouping all Dalbergia as one class. For this latter study, spectra of D. cearensis and D. tucurensis were added to the broader Dalbergia class. These spectra were not included in the first set because the number of specimens was not enough to create an exclusive class for them. PLS-DA presented efficiency rates of over 90% in both situations, while SIMCA presented 52% efficiency at specieslevel separation and 85% efficiency separating D. nigra from other Dalbergia. It was shown that PLS-DA approaches are far better suited than SIMCA for generating a field-deployable NIRS model for discriminating these Dalbergia.