Radon 222 distributions during the 1993 North Atlantic Regional Experiment are simulated with the National Oceanic and Atmospheric Administration Aeronomy Laboratory three‐dimensional regional transport model. The model domain covers the western part of the North Atlantic Ocean and part of the eastern United States and Canada. It horizontally elongates 4200 km in the east‐west and north‐south directions and vertically extends from the surface to 50 mbar. The model simulates emissions, transport, and radioactive decay of 222Rn. Advection, cloud transport, mixing in the planetary boundary layer, and diffusion are simulated and the emissions, boundary, and initial conditions are parameterized. The model results are analyzed to study the ability of the model to simulate 222Rn distribution in the troposphere. The analysis indicates that 222Rn simulations near the surface are dominated by emissions. At altitudes from <1 km to ∼6 km the simulations are noticeably influenced by the side boundaries and the influence increases with increasing altitude. Radon 222 simulations in the upper troposphere (z >6 km) are dominated by side boundary conditions and the wind field. The impact of top boundary on 222Rn simulations is negligible provided the top is specified at an altitude well above the tropopause. Cloud mixing is important in the vertical transport of 222Rn from the surface to the free troposphere. Cloud effect is significant in 222Rn distributions at grid columns within or near which there are abundant clouds. The mean effect of cloud mixing cannot be appropriately represented by contrasting large‐scale or long‐term average 222Rn distributions simulated in cases with and without cloud mixing. Surface pressure fluctuation has substantial influence on 222Rn emission and concentration variabilities, but this effect is restricted to altitudes near the surface. The model‐simulated 222Rn concentrations are compared with measurements made onboard the U.S. Department of Energy Gulfstream aircraft based at Halifax, Nova Scotia, to evaluate the quality of the simulations. The comparison indicates that the simulations and observations are usually within a factor of 2 of one another and significantly correlated with each other. The simulations are capable of capturing high 222Rn concentrations observed in continental plumes.