Abstract: To explore the feasibility of predicting the total volatile basic nitrogen in Tan-mutton by hyperspectral imaging technology and find the best prediction model. Hyperspectral images of 200 mutton samples were collected in the wavelength range of 400~1000 nm, the monte carlo method was used to eliminate abnormal samples. Campetitive adaptive reweighted sampling (CARS), uninformative variable elimination (UVE) and successive projections algorithm (SPA) combination methods were used to superimpose the original spectrum and extract the sensitive wave points. Then the near-infrared spectroscopy quantitative model of the total volatile base nitrogen in mutton and the dynamics model of its variation were established. The results showed that the prediction model of the total volatile basic nitrogen was established by partial least squares regression (PLSR).The prediction correlation coefficient R_p was 0.866 and the root mean square error RMSEP was 3.790. At the same time, the near-infrared spectroscopy model was applied to the zero-order reaction kinetics model of the total volatile basic nitrogen with time, and the correlation coefficient of the model was 0.915.The results demonstrated that the near-infrared spectroscopy technique combined with the kinetic model would predict the safe storage time of the total volatile basic nitrogen in Tan-mutton.