Abstract: In this paper, a qualitative discrimination method for the yolk color of pink shell egg using near infrared spectroscopy is proposed. A total of 90 near-infrared spectral data of pink shell eggs were collected, including 63 eggs in the correction set and 27eggs in the prediction set. The yolk color of 90 eggs was measured by the Roche colorimetric fan, and the number of yolk colors greater than or equal to 8 was set to 1, and the number of yolk colors less than 8 was set to 0. The partial least squares method was used for analysis. The number of principal component was first determined approximately by the interactive root mean square error,and then further discriminant analysis was carried out according to the number of principal component, the final result shows that after the first derivative of the initial spectral data, when the principal component number is 3, the correct rate of non-destructive discrimination prediction is 78.26%.