The oxidation stability of diesel, biodiesel and their mixtures is commonly examined by PetroOxy, Rancimat or thermal techniques, but these analytical methods involve complicated sample preparation and the use of many reagents, making them time consuming and quite expensive. Here, we investigate the possibility of using near-infrared (NIR) spectroscopy as a rapid and cost-effective alternative that, in addition, offers the opportunity of real-time on-line monitoring. First, a calibration model was developed from 133 calibration samples whose reference values were obtained using PetroOxy. Next, based on various chemometric parameters, a NIR model for the prediction of oxidation stability was optimised. The calibration samples were equally distributed across the standard range of 18–128 min. The correlation coefficient of 0.9908 for the NIR model was obtained. Finally, the NIR model was validated using 30 industrial samples. A maximum absolute difference of less than 8 min was determined between the reference and NIR models, with good repeatability (RSD 8.59%). Overall, our NIR model appears to have several advantages for use in determining the oxidation stability of diesel, biodiesel and their mixtures.