Objective:Ensure the safety of final fried food in the process of industrial food production.Methods:The acid values, peroxide values and carbonyl values of the frying oil samples collected from a university cafeteria were determined based on the standard methods of China. Their infrared spectra were obtained with a Fourier transform infrared spectrometer (FT-IR) equipped with multi-point attenuated total reflection (ATR). Partial least squares regression (PLS) was employed to build mathematical models for control oil samples, random oil samples, and all samples, respectively.Results:The models had good predictability for the determination of acid values of frying oils, but did not perform well for peroxide values. In addition, the PLS models for control oil samples had better predictability than that for random oil samples.Conclusion:Fourier transform mid-infrared spectroscopy with attenuated total reflection attachment combined with the chemometric method is expected to achieve rapid and simultaneous detection of acid value, peroxide value and carbonyl value of oil used for frying in the industrial production process of fried foods.