To solve the problems of noise and redundancy in spectra regions which partial least squares discriminant analysis (PLSDA) encounters, synergy interval partial least squares discriminant analysis (SiPLSDA) is proposed and is used in the qualitative analysis model of near infrared spectra of pork. With the help of synergy interval partial least squares (SiPLS), SiPLSDA can select the spectral regions where the prediction model is constructed. Antaris II Fourier transform-near infrared (FT-NIR) spectrophotometer is used to obtain near infrared reflectance (NIR) spectra of pork samples in the range of 10000~4000 cm- 1. NIR spectra are preprocessed by standard normal variate (SNV) transformation. A qualitative analysis model of NIR spectra is built by SiPLSDA. The experimental results show that according to the SiPLSDA prediction model, the recognition ratio of pork storage time is up to 93.94 % and it is higher than that of PLSDA.