Abstract: A method for multivariate adulteration detection on an edible oil includes (1) construction of a model: S1, acquiring near-infrared spectra of edible oils; S2, establishing a near-infrared spectral database of the edible oils; S3, establishing a multivariate adulteration detection model for a type of edible oil; and (2) application of the model: acquiring spectra of a sample to be tested according to the near-infrared spectral signal acquisition method in step S1, preprocessing the obtained near-infrared spectra by using the method in step S2 to obtain near-infrared spectral data of the sample, and determining the authenticity of the sample to be tested by using the multivariate adulteration detection model for the edible oil established in step S3. The method is simple and rapid in operation, can effectively and rapidly screen the authenticity of an edible vegetable oil, and has strong practicability.