Supervising wood exploitation can be very challenging due to the existence of many similar species and the reduced number of wood identification experts to meet the demand. There is evidence that valuable endangered wood species are being smuggled disguised as other species. Near infrared spectroscopy (NIRS) and chemometrics has been successfully used to discriminate between Amazonian wood species using high resolution instruments. In this study, a handheld spectrometer was evaluated for the discrimination of six visually similar tropical wood species using PLS-DA. Woods of mahogany (Swietenia macrophylla) and cedar (Cedrela odorata), both high value tropical timber species included in Appendixes II and III of the CITES, respectively; crabwood (Carapa guianensis); cedrinho (Erisma uncinatum); curupixá (Micropholis melinoniana); and jatobá (Hymenea coubaril). The data for model development and validation take into account both laboratory and field measurements. Outlier exclusion was performed based on Hotelling T2, residuals Q and errors in the estimated class values. The efficiency rates were higher than 90% for all species, showing that the handheld NIR combined with PLS-DA succeeded in discriminate between these species. These results stimulate the application of handheld NIR spectrometers in the supervision of wood exploitation, which can contribute to the species preservation.