In an attempt to mitigate the spread of COVID-19, China implemented a series of strict lockdown measures in 2020, such as restricted road usage and reduced population flow, to safeguard public health. The resulting reduction in anthropogenic pollution emissions caused by the lockdown is informative for researchers seeking to explore the relationship between anthropogenic factors and atmospheric pollutants. Therefore, this study explored severe urban air pollutants and their driving factors before and during the lockdown in China. Across 367 Chinese cities, the implementation of strict lockdown measures led to substantial reductions of 22%–55% in major air pollutants, including PM10, PM2.5, NO2, SO2, and CO. However, O3 levels, which increased by approximately 58.0% during the lockdown, exhibited the opposite trend. This was potentially attributed to the consumption of NOx in the photooxidation process. To more thoroughly understand the relationship between air pollutants and their sources, four representative areas exhibiting considerable variations in air pollutant levels were selected: the Beijing–Tianjin–Hebei region (BTH), the Fenwei Plain (FWP), the Yangtze River Delta (YRD), and Southern central China (SCC). The land-use regression (LUR) model was employed for this purpose. Before the lockdown, various factors such as road-related emissions, rural residential areas, and grassland areas accounted for 13.2%–21.7% of PM2.5 and PM10 concentrations in the FWP region. Conversely, in BTH, YRD, and SCC, land-use-related sources such as plowland areas played an increasingly crucial role, contributing 12.9%–33.6% to PM concentrations. Moreover, human spatial activities and land-use types contributed up to 21.0%, 31.7%, and 20.8% to the elevated levels of SO2, NO2, and CO over the selected regions. During the lockdown, the LUR models revealed that changes in PM2.5 and PM10 concentrations were strongly associated with the influences of plowland and grassland. However, for NO2, SO2, and CO, anthropogenic sources such as trucks on the road, industrial emissions, and construction areas were identified as the main driving factors. Additionally, analysis using the generalized additive model revealed that before the lockdown, high humidity levels exceeding 50% were associated with increases in PM2.5 concentrations in the BTH and FWP regions. By contrast, during the lockdown, gas pollutants such as NO2 and SO2 emitted from livelihood-related activities had an effect on PM2.5 levels, contributing 10%–15% across all four regions. Overall, this study provides crucial evidence to inform the formulation of well-informed air quality strategies by considering key influencing factors in the future.