The potential of visible and near infrared (Vis/NIR) spectroscopy was investigated for discriminating the varieties of Nanfeng mandarin fruit nondestructively. The spectra were collected by a spectrophotometer in the wavelength range of 600–1040 nm. Relationship between the spectra and Nafeng mandarin varieties was established using principal component analysis (PCA), supervised independent modeling of class analogy (SIMCA) and backward propagation neural network (BPNN). By comparison the best result was obtained by BPNN with recognition rate of 97.5%. The results suggested Vis/NIR spectroscopy combination with BPNN was a new approach to discriminate of the varieties of Nanfeng mandarin fruit nondestructively.