Urban parks play an essential role in mitigating the effects of air pollution on human health in a healthy city construction process. However, due to the data limitations, little is known about the spatial distribution of real-time expressed air pollution-related health (APRH) across different urban parks and the contribution of the associated factors. To fill this research gap, this research was conducted based on social media Weibo data (Chinese Twitter) and other geographical data using semantic analyses and the Geo-Detector method by taking 169 urban parks in Beijing as the study area. The results showed that there were more Weibo items relating to APRH clustered within the third ring road and decreasing outward along the ring road. A total of 16 factors in three categories were introduced to analyze the driving forces of this spatial distribution. Accessibility was outstanding with a q-value of the number of subway stations (X14) as high as 0.79, followed by built environment and finally park attributes. Distinguished from those reports based on the traditional statistical data, this research demonstrated that although the urban parks improved the APRH, the exposure to air pollution also increased the health risks when visiting the urban park. It also provides a geographical understanding of the urban parks’ effect on APRH and theoretical guidance for urban park planning and construction.