The liquid N2O concentration data measured by liquid sensors placed in aerated tanks are an input to gas-liquid mass transfer models for the prediction of N2O off-gas emissions. The prediction of N2O emissions from Water Resource Recovery Facilities (WRRFs) was evaluated by three different mass transfer models with the Benchmark Simulation Model 1 (BSM1) as reference model. Inappropriate selection of mass transfer model may result in miscalculation of carbon footprints based on soluble N2O online measurements. The film theory, widely adopted in process modeling, considers a constant mass transfer expression. However, more complex models suggest that emissions are affected by the aeration type, efficiency, and tank design characteristics. The differences among model predictions were 10 -16% at dissolved oxygen (DO) concentration 0.6 g/m 3 , when biological N2O production was the highest, while the flux of N2O was between 20.0 to 24 kg N2O-N/d. At lower DO, nitrification rate was low, while at DO higher than 2 g/m 3 , the N2O production was inhibited leading to higher rates of complete nitrification and a flux of 5 kg N2O-N/d by all models. The differences increased to 14 -26% at deeper tanks, due to the pressure assumed in the tanks. The effect of aeration efficiency and wastewater matrix effect was studied through the correction factor for oxygen transfer (alpha) and oxygen concentration of saturation (beta). The predicted emissions were affected by the aeration efficiency when kLaN2O depends on the airflow instead of the kLaO2.The wastewater matrix had a milder effect. Increasing the nitrogen loading rate under DO concentration of 0.50 -0.65 g/m 3 increased the differences in predictions by 10 -20% in both alpha 0.6 and 1.2. A sensitivity analysis indicated that at this operational point, the mass transfer models did not affect the selection of biochemical parameters for N2O model calibration.