Simulations are commonly used to estimate the effects of air leakage on building energy use. Simulation codes use numerous parameters, such as flow coefficient, pressure exponent, and shelter factor, to estimate the air leakage rate of the building envelope. Input values for these parameters, however, rely on the judgment of the user, and information is not available on the impact of these selections. Oak Ridge National Laboratory partnered with 3M to evaluate how air leakage rate estimates from the popular whole-building energy simulation software EnergyPlus compare to measured values. To this end, we conducted experiments in eight one-story test facilities in Cottage Grove, MN, which were relatively airtight compared with common buildings. The unoccupied buildings were instrumented to gather temperature and heat flux measurements throughout their enclosures, and a weather station was installed at the site to monitor environmental parameters. We used these data to calibrate EnergyPlus models of the test facilities. Blower door tests were performed to measure air leakage rates at certain pressure ranges to develop correlations between the air leakage rate and the indoor-outdoor pressure difference. Tracer gas tests were conducted to measure air leakage rates at normal operating conditions. Flow coefficients and pressure exponents calculated from blower door test results and various user-selected input parameters were used with the calibrated EnergyPlus models to predict the air leakage rates. EnergyPlus estimates were compared to the measurements from the tracer gas tests. Findings include that EnergyPlus air leakage estimates that were calculated using flow coefficients and pressure exponents from pressurization tests yield results that are closer to those from the tracer gas tests than data from the depressurization tests. Results from this study provide guidance on the parameters that improve EnergyPlus air leakage rate estimates, which will lead to more realistic energy-savings predictions.