Abstract: The superior varieties breeding for navel orange is completed by the selection of bud mutation, and it is necessary to identify navel orange varieties in order to select some specialized character varieties for the cultivation. In this study, near infrared spectroscopy (NIRS) coupled with Soft Independent modeling of class analogy (SIMCA) and Partial Least Squares - Discriminant Analysis (PLS-DA) pattern recognition methods were used to identify the navel orange varieties. The results showed that four navel orange varieties were well identified by raw NIR spectra and SIMCA models, the identification rates were all 100%. For PLS-DA models, the correlations between the predicted category variables of calibration or validation and the measured category variables were all remarkable with a correlation coefficient (r) over 0.970 and low RMSECV and RMSEP (＜0.100); The discrimination accuracy for the navel orange varieties was 100% by PLS-DA model based on the validation set of samples. It is suggested that Vis/NIR spectroscopy coupled with SIMCA and PLS-DA methods can be used for rapid detection of navel orange in superior varieties breeding.