Environmental problems driven by urbanization threaten public health and urban vitality. Improving urban wind environments helps regulate local climate, disperse air pollutants, and enhance quality. Traffic-oriented pocket parks support daily public activities, yet their wind environments are affected by vehicle emissions and complex airflow. However, studies on pedestrian-level wind environments in pocket parks are limited, and integrated analyses linking environmental performance with human activity are insufficient. This study proposes a downscaling simulation framework coupling the mesoscale Weather Research and Forecasting (WRF) model with the microscale ENVI-met model to simulate pedestrian-level wind environments under a unified meteorological background. Typical traffic-oriented pocket parks in Nanjing were selected as case studies. Wind conditions in winter and summer were evaluated, and simulation results were overlaid with the spatial distribution of crowd concentration to identify three spatial types: core conflict zones, seasonally adaptive zones, and stable lowresponse zones. Results show that the coupled WRF-ENVI-met approach overcomes limitations of traditional field measurements and enables comparative wind environment assessment across multiple pocket parks. Seasonal differences in wind conditions were observed, and pavilion and corridor structures influenced local airflow. Spatial correspondence was identified between wind environment quality and crowd concentration, though this relationship is affected by multiple factors and requires further validation. This study supports climate-adaptive green space design in high-density urban areas.