Objective To study the rapid identification of freezing-thawing meat by stoichiometry combined with low field nuclear magnetic resonance（LF-NMR）.Methods CPMG sequences were selected by LF-NMR to collect 111 NMR signals from pork tenderloin. Through the inversion of NMR signal data， four principal components were extracted from 12 variables by principal component analysis， and the discriminant research model was established.Results According to the analysis of fresh meat， slightly freezing-thawing meat and freezing-thawing meat， the accuracy of the model was 95.5% verified by back generation method， and it was 94.6% by cross-verification.Conclusion LF-NMR is simple and rapid， and can be used for discriminant analysis of freezing-thawing meat， which provides a reliable and effective basis for the supervision department.