Abstract: Based on the MODIS aerosol optical depth (AOD) products and monitoring data from township ambient air automatic monitoring stations, the spatiotemporal variation patterns and spatial autocorrelation of AOD in Shijiazhuang were quantitatively analyzed from a seasonal perspective. A geographically and temporally weighted regression (GTWR) model was utilized to invert atmospheric pollution parameters, Research on Air Pollution. The main conclusions are as follows: ① The seasonal spatial distribution of AOD exhibited distinct characteristics. Influenced by regional emission sources and meteorological conditions, the seasonal mean values were lower in spring and summer, and higher in autumn and winter. The overall spatial distribution showed low-value aggregation areas in the northwest and high-value aggregation areas in the southeast. ② The seasonal spatial distribution of AOD demonstrated a high degree of spatial autocorrelation. Its spatial aggregation characteristics could effectively represent the spatial differentiation patterns of atmospheric pollution. Particularly in areas lacking ground-based monitoring data, AOD can serve as an important indicator for dynamic pollution monitoring. ③ The combined use of AOD and the GTWR model efficiently captured the spatiotemporal dynamic characteristics of atmospheric pollution, revealing the spatial heterogeneity in the relationship between AOD and pollution factors. This provides a scientific basis for regional pollution source apportionment and the formulation of prevention and control strategies.