AbstractRapid urbanization adversely impacts urban air quality, impacting human health and the ecological environment. However, few studies have systematically explored air quality from the perspective of urban structural evolution. This can be done using two scales: one focusing on microscopic patches and one focusing on the macroscopic built-up area. Based on polycentricity at a city scale, urban growth patterns at a patch scale, and other natural and socioeconomic data for 362 cities in China, this study applied spatial regression models to quantify the impact of urban structure on PM2.5. The PM2.5 data were downloaded from the NASA Earth Observing System Data and Information System (EOSDIS) data center. The results show the following: (1) at a city scale, the polycentricity index had a significant and negative relationship with PM2.5, the development of polycentric cities supports improvements in air quality; (2) at a patch scale, three urban growth patterns (edge-expansion, outlying, and infilling) all significantly and negatively impact PM2.5. Urban growth continues to increase PM2.5 pollution in China; and (3) road density, the built-up area, population density, land-use mix, and other control variables have an influence on the PM2.5 concentration, and this study verified the existence of a scale effect. In addition, considering the different urban development trajectories of developed and developing countries, this study makes targeted recommendations for urban managers wanting to make reasonable development policies. In general, quantifying the relationship between urban structure and PM2.5 helps provide a scientific basis for city planners and managers to optimize urban layouts.