Abstract: In order to discriminate the rice flour processed by different doses of irradiation, a fast and nondestructive method was proposed based on the visible-near infrared spectroscopy. Four groups of rice flour were irradiated using different doses of 60Co γ-rays, and 200 test samples were obtained. The reflection spectrum data of all samples were collected by using ASD visible-near infrared spectrometer, and the data were analyzed by principal component analysis (PCA) method. Taking the first 6 principal components (PCs) as the inputs of the back-propagation artificial neural network (BP-ANN), an identification model was established. The results showed that the identification accuracy of the model for predicting samples could reach up to 100% in the setting of standard deviation of ±0.1. The proposed method has good classification and identification effects, which can provide a new technical method for fast identification of irradiation rice flour products.