In this paper, the effect of temperature on the near-infrared spectroscopy model of cement raw meal was investigated by developing two models (model I and model II) with different calibration sets. Model I and model II were developed by applying the partial least squares method using same-temperature and different-temperature samples, respectively. In comparison with model I, the root mean square error of prediction (RMSEP) of SiO₂, AlO₃, Fe₂O₃, and CaCO₃ in model II was reduced by 78.3%, 26.4%, 42.9%, and 60.4%, respectively. Experimental results show that the temperature of the cement raw meal sample has a certain influence on the prediction results of the near-infrared spectroscopy model. The influence of temperature on the prediction results can be effectively reduced by modeling with temperature gradient samples in the calibration set, and thus the infrared spectroscopy model can better be applied to production site.