In this work, gas chromatography-ion mobility spectrometry (GC-IMS) was used to analyze the volatile organic compound changes of rapeseed oil with different refined grades, the odor fingerprints of refined rapeseed oil were constructed, and a nonlinear model was built to realize rapid and accurate discrimination of rapeseed oil with different refined grades. 124 rapeseed oil samples with different refined grades were collected and analyzed by GC-IMS and chemometric tools, and 34 characteristic peaks were selected by the colorized difference method as variables to characterize the internal quality in rapeseed oil of different refined grades. The principal component analysis algorithm was used to further reduce dimensionality and extract the most relevant information. The -nearest neighbor algorithm was applied to build a discriminant model. All the samples were recognized accurately without errors, and the results show the potential of this method to discriminate different refined grades of vegetable oil.