The frying qualities of palm and soybean oils are determined using infrared spectroscopy and multivariate calibration. Compare to soybean oil, palm oil is more resistive to the chemical and physical changes and this is attributed to the high degree of unsaturation of soybean oil (61.9%) compare to palm oil (13.8%). After 48 h in service, the oil samples were effectively clustered into two groups using principal component analysis which indicated that both oils still maintain their chemical identities. Partial least squares regression (PLS1 and PLS2) a long with mid-FTIR data are used for predicting free fatty acid, viscosity, and total polar compounds of the used oils without running expensive standard procedures. PLS1 and PLS2 outperformed PCR and MLR for predicting the quality indicators of the frying oils. For palm oil and at the optimum calibration conditions, the obtained accuracies (SD) are 105.6% (0.05), 99.8% (1.10), and 103.9% (0.16) for free fatty acid, viscosity, and total polar compounds, respectively. The proposed method is simple, less-expensive, and has comparable accuracy/precision with standard procedures that often used for monitoring frying oils.