Abstract: To investigate the quality alteration of walnut oil during oxidation and to achieve the rapid discrimination of the oxidation degree of walnut oil, the experiment was conducted to analyze the quality changes of walnut oil under accelerated oxidation conditions using walnut oil as the raw material with different oxidation degrees; and the olfactory analyzer was used to explore the volatile odor variation of walnut oil with different oxidation times by integrating radar chart, principal components analysis (PCA), orthogonal partial least squares discriminant analysis (OPLS-DA), and correlation analysis. Simultaneously, the response values of the olfactory analyzer and the peroxide value and acid value were utilized to establish a prediction model for the rapid discrimination of the oxidation degree of walnut oil through partial least squares regression (PLSR) and support vector regression (SVR) approaches. The results indicated that the peroxide value, acid value, and anisidine value reached 6.49 g/100 g, 1.11 mg/g, and 187.41 respectively after 48 hours. Meanwhile, total tocopherols decreased from 11563.33 μg/100 g to 546.71 μg/100 g by the end of the oxidation. Peroxide value, acid value, and anisidine value all showed a positive correlation with a significant difference level of P<0.001; total tocopherols showed a negative correlation with a significant difference level of P<0.001 with peroxide value and anisidine value; and negative correlation with significant difference level of P<0.05 with acid value. The olfactory analyzer could effectively distinguish the volatile compounds of walnut oil and the changes in the response values of the olfactory analyzer could determine the oxidation degree of walnut oil. The PLSR model corresponding to the peroxide value had a high predicted coefficient of determination (Rp) of 0.9361, suggesting that the PLSR method was the most suitable in the prediction model established by the peroxide value and the response value of the olfactory analyzer. Conversely, the SVR method performed more conspicuously in the model established by the acid price and the response value of the olfactory analyzer (with a Rp of 0.9042), and both models functioned exceptionally well in predicting the degree of oxidation of walnut oil, which could offer a scientific foundation for assessing the degree of walnut oil oxidation.