Microscale air quality in highly populated urban areas has gained increasing attention in recent years. Various emission sources are present, and their contributions need to be quantified for assessing human exposure and developing effective emission control strategies. This dissertation presents the effort towards establishing a better understanding of the spatial variation of multiple air pollutants in complex urban microenvironments through numerical modeling and experimental evaluation. In the first part, I investigated the transport of Black Carbon (BC) in a typical highwaybuilding environment next to an urban high school in South Bronx, NYC. Two generalized configurations i.e., highway-building canyon and highway viaductbuilding are discovered, which is critical to the spatial variation of BC. The second part focuses on roadside barrier designs with the objective to mitigate near-road air pollution. Our analysis revealed two potentially viable design options: a) wide vegetation barriers with high leaf area density which reduces downwind particle concentrations significantly, while resulting in a moderate increase in on-road concentrations, and b) vegetation-solid barrier combinations lead to the greatest reduction in downwind particle concentrations among all configurations and a large increase in on-road concentrations at the same time. The third part investigates the near-source environmental impact of diesel backup generators that participate in demand response programs. The micro-environmental air quality simulation is improved by coupling with a meteorology processor to provide realistic boundary conditions. The study found the near-ground PM2.5 concentration for the worst scenarios could well exceed 100 [MICRO SIGN]g m-3, posing a potential health risk to people living and working nearby. Our analysis also implies that the siting of diesel backup generators stacks should consider not only the interactions of fresh air intake and exhaust outlet, but also the dispersion of exhaust plumes in the surrounding environment. The last part studies the environmental impact of a biomass boiler with and without PM emission control. A micro-environmental model was applied to simulate the experimental conditions, and a good agreement between predicted and on-site measurement is observed. Our analysis shows that the absence of ESP could lead to an almost 7 times increase of the near-ground PM2.5 concentrations in the surrounding environment.