Attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy-based prediction methods were previously shown to be a practical alternative to determine milk adulteration. However, milk in general contains high amount of water and infrared absorptions sourcing from water dominate the mid-infrared spectrum of the milk, preventing the other key vibrations in fingerprint region from escalating. Therefore, we aimed at evaluating the efficacy of a 30-s vacuum treatment on liquid milk aliquots prior to spectral collection and compare the results with no vacuum-applied spectra. Overall, quick vacuum treatment approach led to better-resolved peaks in the fingerprint region of the spectra compared with water-dominated spectra of milk with no vacuum application. Considering the residual predictive deviation (RPD) values obtained, vacuum treatment has significantly improved the performance of the models in prediction indicating that quick vacuum approach would be beneficial if implemented for routine determination for milk adulteration.