Near-infrared (NIR) spectroscopy was investigated as a method for analysis of the carbonyl value (CV) in frying oils. Two types of rapeseed oil were used to fry potatoes and 156 samples of frying oil were obtained. Transmission spectra of the samples were acquired directly using glass test tubes (13-mm diameter) as sample cells and an NIR spectrometer. Calibration models with very high accuracy based on the second derivative spectra in wavelength range of 1100-2200 nm were developed for predicting the CV using partial least-squares regression with full cross-validation. The coefficients of determination for calibration and standard error of cross-validation were 0.98 and 1.79 μmol g −1 , respectively. The accuracy of the NIR calibration models was tested using the validation set, yielding the value for the root mean square of the prediction as 1.74 μmol g −1 . The results demonstrate that the CV for frying oils can be accurately determined using NIR spectroscopy.