It’s a common phenomenon that shoddy and deceptive the hongmu furniture is filled with all markets.It’s so hard for consumers to identify the species of the hongmu furniture.There are many studies on the application of near infrared spectroscopy classification of wood species.But for the finished furniture, the surface decoration is the biggest problem to classify the hongmu furniture.So this paper analyzes the spectral characteristics of wood covered by different kinds of paints and proposes a method which is combined with Principal Component Analysis(PCA) and Back Propagation Neural Networks by using near infrared spectroscopy.We firstly collected spectral data from Dalbergia cochinchinensis and Dalbergia retusa which are usuaslly confusing hongmu furniture species in the market, and compered the spectral difference of hongmu and the hongmu furniture.Secondly, reflection spectra were preprocessed by taking SNV to remove irrelevant information and noise.Finally, principal component analysis was used to distill initial feature vectors and asbilish new sample’s feature vectors according to cumulate reliability.Then, we chose the new feature vectors which its cumulate reliability is more than 95%.A three layer probabilistic neural network was designed.Results show that recognition rates of these three species of hongmu furniture are 92.59% and 94.38%.So our approach provides an new and reliable method for the identification of hongmu furniture.