In order to achieve rapid nondestructive determination of phosphate content in hairtail surimi and surimi-based food(kamaboko), near-infrared quantitative models were established by fitting the near-infrared diffuse reflectance spectral data of haitrail surimi and kamaboko with its phosphate content. Prediction and correlationcoefficients of thecalibration and validation were used to evaluate the quality of the model. The model was further developed by partial least squares(PLS). The result showed that correlationcoefficients calibration in the model of hairtail surimi and surimi-based food(kamaboko) were 0.983 and 0.960, respectively and prediction of calibration were 0.032 and 0.101, respectively. The correlationcoefficients of thevalidationswere 0.951 and 0.954, respectively. The predictions of validation were 0.058 and 0.097, respectively. This study demonstrated that NIR spectroscopy could be used to determine the content of phosphate in hairtail surimi and surimi-based food(kamaboko) to a certain extent.