The invention provides a method for identifying organic apples and common apples based on NMR metabonomics. The method comprises the following steps: obtaining NMR fingerprints of organic apples and common apples by using a nuclear magnetic resonance technology, and analyzing the fingerprints to identify chemical compositions of the organic apples and the common apples; analyzing the chemical compositions of the organic apples and the common apples by using a supervised multidimensional statistical analysis method, namely, orthogonal partial least squares discriminant analysis (OPLS-DA); validating the reliability of the established model by two validation methods of internal cross-validation and CV-ANOVA in external validation in the OPLS-DA model, and finally solving the technical problem of identifying the organic apples and the common apples by a Score graph in the OPLS-DA model.