Abstract: In order to discriminate the floral origins of honey,the concentrations of 13 mineral elements( K,Na,Zn,Mn,Mg,As,Fe,Cr,Ni,Ca,Cu,Pb,and Cd) of three honeys( Vicia cracca honey,Hevea brasiliensis honey and Punica granatum honey)from Yunnan( China) were determined by flame atomic absorption spectrometry( F- AAS) or graphite furnace atomic absorption spectrometry( GF- AAS),which showed great differences among the honeys. Based on the specific mineral content,principal component analysis( PCA),partial least- squares discriminant analysis( PLS- DA) and back- propagation artificial neural network( BP- ANN) were used in classification of the three honeys. With PCA,three honey species were preliminary classified by three principal components,which were established from thirteen mineral contents. Subsequently,PLS- DA and BP- ANN classification model were constructed with 30 randomly selected samples from the three honey species. In PLS- DA,the total correct classification rates for model training and cross- validation were 96.7% and 92.2%,respectively.In BP- ANN,the total correct classification rates for model training and cross- validation were 100% and 95.6%,respectively,indicating a better performance of BP- ANN than PLS- DA. The validation of BP- ANN model was further tested by the rest 35 honey samples.H.brasiliensis honey and P.granatum honey samples were predicted with 100% accuracy.V.cracca honey samples was predicted with 90% accuracy.These result suggested that the value of mineral content tested by F- AAS or GF- AAS with chemometric methods could be used as a potential and powerful tool for the classification of honeys from different botanical origins.