The development of rapid and simple test is needed in order to capably monitor the frying oil quality and ensure the fried food safety.In this paper，the evolution of several analytical indicators（e.g.acid value，viscosity，absorbance values and TPC content） and the low field Nuclear Magnetic Resonance（LF-NMR） T2 relaxation characteristics （the relaxation time:T21，T22，T23，the corresponding peak area:S21，S22，S23 and the single component relaxation time:T2W） of soybean oil were studied during the frying process of no foodstuff or potato chips（180±5 ℃ for 36 h）.The multiple regression analysis method was used to establish the correlation model between LF-NMR results and the analytical indicators.At last，the developed models were experimentally verified.The results indicated that the acid value，TPC content and S21 increased linearly，while T21，T22 and T2W decreased linearly，as the frying process prolonged.And a good binomial relationship was obtained between the viscosity or absorbance values and the frying time（r2＞0.90）.No obvious relationship was found between frying time and T23，S22 or S23 .Compared with no foodstuff frying，the acid value，viscosity，TPC content，absorbance values and S21 during potato chip frying process increased significantly，while T21，T22 and T2W obviously shortened （P<0.05）.Multiple regression analysis showed that both acid value and TPC content had a good relationship with T2W and T21，and viscosity had a good linear relationship with T2W，while absorbance had a good logarithmic relationship with S21（r2＞0.93）.These models were successfully verified and could be used to predict the variations of physical and chemical indicators by the LF-NMR characteristics of the frying oil.