This paper presents a computational fluid dynamics (CFD) simulation approach for the prediction of leakage between an N95 filtering facepiece respirator (FFR) and a headform and an infrared camera (IRC) method for validating the CFD approach. The CFD method was used to calculate leak location(s) and 'filter-to-faceseal leakage' (FTFL) ratio for 10 headforms and 6 FFRs.The computational geometry and leak gaps were determined from analysis of the contact simulation results between each headform-N95 FFR combination. The volumetric mesh was formed using a mesh generation method developed by the authors. The breathing cycle was described as a time-dependent profile of the air velocity through the nostril. Breathing air passes through both the FFR filter medium and the leak gaps. These leak gaps are the areas failing to achieve a seal around the circumference of the FFR. The CFD approach was validated by comparing facial temperatures and leak sites from IRC measurements with eight human subjects. Most leaks appear at the regions of the nose (40%) and right (26%) and left cheek (26%) sites. The results also showed that, with N95 FFR (no exhalation valves) use, there was an increase in the skin temperature at the region near the lip, which may be related to thermal discomfort. The breathing velocity and the viscous resistance coefficient of the FFR filter medium directly impacted the FTFL ratio, while the freestream flow did not show any impact on the FTFL ratio. The proposed CFD approach is a promising alternative method to study FFR leakage if limitations can be overcome.