In order to realize the rapid identification of different varieties of pears, principal component analysis (PCA) on the spectral data clustering analysis was used on three different varieties of pears to find the characteristic differences. The principal component analysis showed that the main composition PC1 and PC2 for all the modeling samples score diagrams had very good clustering effect to the different types of pears. Load diagram that got by using principal component analysis can obtain the variety sensitive characteristic wavelengths from pears, and with the characteristic band spectrum as input to build partial least-squares discriminant (DPLS) and least squares support vector machine (LS—SVM) models. Seventy pears of three varieties with 210 in total were used to build DPLS and LS—SVM models respectively. The unknown 24 samples were predicted by the models, the recognition accuracy rate of the LS—SVM model reached to 100%. The calibration and verification results of the DPLS model and the actual classification variables of the correlation coefficient was greater than 0.980. Cross validation root mean square error (RMSECV) and root mean square error of prediction (RMSEP) were less than 0.100. The varieties recognition rate was 100%. The proposed rapid identification method has good classification effects. 