This paper proposes a computational fluid dynamics (CFD) model, along with dimensionless quantitative assessment standard—air pollution residual time (APRT) for the evaluation of local haze-fog (HF) dispersion in a built environment. A low APRT value ensures good ventilation. A building group model that comprises high-rise business building, mid-rise office buildings, low-mid-rise residential buildings (at the center of the building group), a mid-rise recreational center, and a local road (open terrain), was scaled down (1:100) to simulate the HF dispersion process. The orientation of the building group was numerically modified to generate a wind incidence normal to the high-rise building, mid-rise buildings, recreational center, and road. The results showed that the orientation of the building group largely determines the APRT. The most favorable orientation can reduce APRT by more than 50％. Our results strongly suggested that in order to reduce the consequential negative effect of air pollution, future urban designs should undergo a comprehensive ventilation assessment to ensure a low APRT value.