A method for predicting the quality change of frozen pork during storage is provided. The method features that low field nuclear magnetic resonance (LFNMR) and meat quality index measurement are usedas main measuring tools, and the transverse relaxation time T2 spectrum data, TVB-N, TBARS, protein solubility, juice loss rate, and total amino acid loss rate are used to establish a principal component analysis model; the method then establishes a regression model between the comprehensive score and time. Based on the measured data of pork frozen for 0, 1, 3, 6, 9 and 12 months, the regression model was solved, and the accuracy of the model is verified by prediction. As the analysis process is simple, the method can be used to accurately predict the remaining shelf life of frozen pork, to master and understand the nutritional status and freshness of frozen pork, and to provide a theoretical reference for the research of other frozen meat freshness and shelf life prediction models.