In order to identify honeys according to their floral origin, inductively coupled plasma mass spectrometry (ICP-MS) combined with principal component analysis (PCA) and discriminant analysis (DA) were employed in the present study. Three kinds of honeys such as acacia honey samples, sunflower honey samples and rape honey samples were selected. It was pretreated by wet-acid digestionand measured 20 kinds of mineral elements in honey samples by ICP-MS. The result showed that the accuracy of the inductively coupled plasma mass spectrometrymeted the requirements. The result of principal component analysis demonstrated that the acacia honey samples were performed a trend of certain gather. The trend of the sunflower honey samples and the rape honey samples are not obvious. Ten kinds of mineral elements including Na, Mg, K, Ca, Sr, Ba, V, Fe, Ni, Sb can be regarded as honey varieties of characteristic elements. Seven kinds of mineral elements such as Mg, Sr, Ba, Sb, Ni, Cr and Na could be selected through stepwise discriminant analysis. Using bayes discriminant analysis, A linear discriminant function can be recieved. The discrimination rate of honey samples such as acacia honey samples, sunflower samples and rape honey samples were 100%, 80% and 90.9% respectively. Two sunflower honey samples was misclassified into rape honey samples andone rape honey samples are also misclassified into acacia honey sample. The total rate of discriminant model cross validation was 90.3%. It is concluded that the mineral elements in honey varieties with good classification. The present study can provide theoretical basis and the relationship between thetypes ofhoney samples with mineral elements. The method what this study used had simple, accurate and stablecharacteristics, which can be used as a reliable method of honey sample identification.