For the rapid identification of common plant honey and Chinese medicine nectar plant honey,a novel qualitative identification method combined Kernel Principal Component Analysis （KPCA）and Least Square Support Vector Machine （LSSVM）is proposed.The presented method uses Fourier transform NIR spectrophotometer to collect the spectral data of locust,leonuri and coptis honey.The KPCA algorithm is used to extract nonlinear features.Then a LSSVM classificaion model based on Error Correcting Output Code （ECOC）is designed and grid search method is used to determine the optimal model parameters.To unknown honey samples,the optimal model has the best identification capability with a accuracy of 96.67％.Experimental results indicate that the proposed qualitative analysis method based on KPCA and LSSVM can distinguish the honey of common and Chinese medicine nectar plant.