The postmortem diagnosis of fatal heatstroke (hyperthermia) is a challenging task in forensic practice. Attenuated Total Reflectance Fourier transform infrared spectroscopy combined with chemometrics was utilized to explore the biochemical difference of fatal heatstroke and to establish diagnosis models in this study. The complementary results of spectral absorbance comparison and principal component analysis demonstrated that the spectral differences between the hyperthermia and control groups mainly arose from the variations in the structure or composition of nucleic acids, proteins, amino acids, and carbohydrates in the vitreous humor. A genetic algorithm-guided partial least squares discriminant analysis model was established for the postmortem diagnosis of heatstroke, which achieved great specificity and sensitivity, both with scores of 100%. Furthermore, classification predictions of unknown vitreous humor samples collected from 11 rabbits were also performed by the model, resulting in 100% correct discrimination. This preliminary research demonstrates that Attenuated Total Reflectance Fourier transform infrared spectroscopy in combination with chemometrics has the potential to be an effective aid for the postmortem diagnosis of fatal heatstroke.