The present work quantifies numerically the systematic errors present in experimental infrared heat flux studies of boiling surfaces. A transient conduction model for multilayer structures is proposed to describe the periodic heat fluxes encountered on boiling surfaces. The results of the current work show that the systematic error behavior of the infrared method is not uniform but dependent on the frequency of the heat flux signal of the boiling surface; which is a novel finding. As the frequency of the heat flux signal increases, the errors in the measured phase of heat flux signals are expected to increase. The errors in the amplitude of heat flux signals sharply increase at low frequencies (1–10 Hz) and decrease as the frequency increases. The maximum errors in the phase and amplitude of heat flux signals are 9% and 23%, respectively in the frequency range of nucleate boiling (10–80 Hz). Based on the current analysis, it is concluded that the systematic errors found arise from assuming that thermal contact resistances of such systems are negligible. This is an assumption universally adopted by the field. By considering and correcting for the thermal contact resistance in the measurement of heat fluxes, the maximum errors in the phase and the amplitude of heat flux signals can be reduced to 7% and 9%, respectively. The results are applied to experimental data ensembles from the published public domain. Finally, the current work provides general guidelines to improve systematic errors in the measurement of heat flux for the study of boiling using infrared thermography found in the literature.