Abstract: The objectives of this research was to develop a hyperspectral imaging system to predict pork freshness based on quality attributes such as total volatile basic nitrogen (TVB-N) and pH value. Reflectance spectra were collected from the hyperspectral scattering images in the range of 470 to 1 000 nm, and pre-processed by Savitzky-Golay (S-G) based on five and eleven smoothening points and multiple scattering correlation (MSC) methods separately. Their prediction results were compared with prediction models developed by partial least square regression (PLSR) method. PLSR with S-G pre-processing could predict pork TVB-N with correlation coefficient (Rv) of 0.90 and standard error of prediction (SEP) of 7.80. Similarly PLSR with MSC pre-processing data predicted pork TVB-N with Rv of 0.89 and SEP of 8.0. The prediction model established using MSC as pre-processing method yielded better result for prediction of pH value, which Rv was 0.79 and SEP was 0.37. The result showed that, by the prediction models for TVB-N and pH value with MSC pre-processing method, the prediction accuracy for pork freshness classification could reach up to 91%. The research demonstrates that the hyperspectral imaging technique can be applied in rapid and non-destructive detection of pork freshness.