Two modern office buildings in Shanghai with typical variable air volume ( VAV ) systems were selected. Four separate spaces on a standard office floor in each building facing in different directions were selected for site measurements of outdoor airflow rates and indoor air quality (concentrations of CO2 and PM10) during typical days of the four seasons. Computer simulations and calculations were also made of outdoor airflow rates and CO2 concentrations in the four-test spaces on an hourly basis over the whole year. Monthly electrical consumption for the two buildings was recorded or estimated. Simulations and calculations were performed of the buildings' energy consumption and energy cost using two different outdoor air control strategies of a typical VAV system as well as a fan coil unit ( FCU ) system. The data and simulation results were compared and analysed. It was shown that in a VAV system, the outdoor airflow rate distributed to each zone varies greatly, especially during part-load hours, making it difficult to always ensure sufficient outdoor air in each zone and avoid indoor air quality (IAQ) problems. However, this problem can be prevented by using appropriate outdoor air control strategies - e.g. a fixed high level total outdoor airflow rate. 15 refs.