As vehicle population and activity increase, vehicle emissions are becoming the most predominant source of air pollution in Shanghai, China. It has become important to accurately estimate the traffic emissions in this city. This paper presents a bottom-up approach based on the International Vehicle Emission (IVE) model to develop the vehicle emission inventory for Shanghai. The results show that the total emissions of CO, VOC, NO(X) and PM from vehicles in Shanghai in 2004 were 57.06 x 10(4) t, 7.75 x 10(4) t, 9.20 x 10(4) t and 0.26 x 10(4) t, respectively. About 20% of the total emissions were emitted during the cold start period. Heavy-duty vehicles such as trucks and buses contributed over half of NO(X) and PM. Motorcycles and mopeds provided 45.0% of VOC and 36.3% of PM. Light-duty vehicles are the main source of CO emissions. An assessment of vehicle emissions by time of day and road type was also discussed. The three peak emission periods accounted for 54% to 56% of the total emissions during the day and more than 50% of the total emissions were emitted on the arterial roads. Finally, the study focused on the uncertainty analysis of two critical factors: emission factors and the estimate of the total Vehicle Kilometers Traveled (VKT). The analysis indicates that the emission factors calculated in this paper are close to those factors measured during on-road testing, and the difference between the VKT used in this paper and other calculations is less than 10%.