Multispectral radiation thermometry (MRT) was applied to predict the aluminum surface temperature. Experiments were conducted to measure the spectral intensity values for five different aluminum alloys, AL1100, AL2024, AL5083, AL6061, and AL7005, at 600 K, 700 K, and 800 K. The experimental work is coupled with six MRT emissivity models encompassing mathematical and analytical functions to infer surface temperature. Assessment of the MRT emissivity model is subject to parametric effects of number of wavelengths, alloy composition, and temperature. Results show that increasing wavelength number does not significantly improve measurement accuracy while applying MRT. If the emissivity model can represent well the real emissivity behaviors, a more accurate inferred temperature can be achieved. Overall, most models achieve high accuracy in temperature prediction, except two emissivity models. One particular emissivity model provides the best compensation for the aforementioned parametric influences.