In this paper, the pH value of cold fresh pork was non-destructively detected based on hyperspectral imaging (HSI) technique, and some useful data processing methods were discussed. After some sample set partition methods, some spectral pretreatment methods, and some optimum wavelength selection methods were compared respectively, the most suitable data processing method was chosen and the robust hyperspectral model for predicting the pH value of cold fresh pork was established. The results indicated that the pH value hyperspectral model of cold fresh pork established by using the whole wavelengths after the sample set was divided by using concentration gradient (CG) algorithm, and the spectral data was pretreated by using normalization combined with mean center(MC) had the best prediction abilities, with the determination coefficients

$$\text{R}_{{\text{cv}}}^{\text{2}}$$ Rcv2 equaled to 0.768, $$\text{R}_{\text{p}}^{\text{2}}$$ Rp2 equaled to 0.694, RMSECV equaled to 0.1113, and RMSEP equaled to 0.1204. The results also indicated that the model established by using the characteristic wavelengths which were selected by using CARS algorithm had better prediction abilities, with $$\text{R}_{{\text{cv}}}^{\text{2}}$$ Rcv2 equaled to 0.8581, $$\text{R}_{\text{p}}^{\text{2}}$$ Rp2 equaled to 0.8668, RMSECV equaled to 0.0858, and RMSEP equaled to 0.0772. All the results showed that suitable data processing methods was advantageous to the prediction ability of the model, and that HSI technique can be utilized to measure the pH value of cold fresh pork in a rapid and non-destructive way.