The potential of visible/near-infrared spectroscopy (vis/NIRS) for its ability to nondestructively differentiate rough rice varieties (species and years) was evaluated. Partial least squares (PLS) analysis was performed on the processed spectral data. In terms of the total classification results, the model with the preprocessing of wavelet transformation is the optimal to predict, and its prediction statistical parameters were r p of 0.8888, SEP of 0.8029 and RMSEP of 0.8030. This research shows that vis/NIRS has the potential to be used for the discrimination of rough rice varieties, and a suitable pre-processing method should be selected for spectrum data analysis. Assignment of the bands due to chemical content was proposed based upon the PLS loading weights, and the wavelengths 770nm, 970nm corresponding to water, 880nm to fat, 906nm, 922nm, 972nm, 996nm to protein were found.