ABSTRACT Three-dimensional modeling on the aerodynamics of airflow and diffusion of air pollutants in a longitudinal-ventilated traffic tunnel was carried out. The model takes ventilation fans, traffic flow rate, speed, emission factor and piston effect of moving vehicles into consideration. Turbulent flow and dispersion of gaseous pollutants in road tunnels were solved numerically using the finite volume method. Traffic emissions were accordingly modeled as banded line sources along the tunnel floor. The effects of fan ventilation, roughness and piston effect of moving vehicles on the air flow and pollutant dilution are examined. Concentrations of gaseous pollutants CO, NOX, SO2 and THC (total hydrocarbons) at three axial locations in the tunnel, together with traffic flow rate, traffic speed and types of vehicle were measured. Case study was conducted on the Cross-Harbor Tunnel and the Chungcheng Tunnel in which on-site measurements of traffic flow were also conducted concurrently to provide traffic emission data to the tunnel environment for numerical simulation and comparisons. The aim of this study was to understand the spatial variation of air pollutants generated by traffic emissions and evaluation of ventilation performance and piston effect of moving vehicles on dilution of air pollutants in these tunnels. The results show that the major emission sources of CO are passenger cars and motorcycles, while major emission sources of NOx are trucks. Pollutants convect downstream with the wind generated either by longitudinal ventilation fans and/or moving vehicles, thus causing increasing pollutants concentrations with increasing downstream distance. The piston effect of moving vehicle alone can provide 64% ~ 85% increase of wind speed in Chungchen Tunnel and 13% ~ 20% in Cross-Harbor Tunnel. When all fans are on, showing 185% ~ 328% and 120% ~ 182% increases in Chungchen Tunnel and Cross-Harbor Tunnel, respectively. The piston effect of moving vehicle alone can provide 14% ~ 32% dilution of air pollutants in the Chungcheng Tunnel. The piston effect of moving vehicles is compounded with ventilation fans, showing a 47% ~ 66% dilution effect when all fans are on. For the Cross-Harbor Tunnel, the piston effect of moving vehicle alone can provide 9% ~ 23% dilution of air pollutants and 36% ~ 74% dilution effect when all fans are on. The results reveal that cross-sectional concentrations are non-uniformly distributed and that concentrations rise with downstream distance. When all fans were turned off, wind speed in tunnels would be considered as constant, and gaseous pollutants concentration agree with linearly alone the tunnel.