Despite the widely acknowledged benefits of urban green space on human health, few studies have holistically investigated residents of urban areas in China. Relying on citywide cross-sectional survey data (2013–2015, n = 2,648), this study examines how green space affects general health conditions gauged by the self-reported health (SRH) of the urban population in Beijing. Two indicators are applied to measure the accessibility of green space: one uses 300-m, 500-m and 1,000-m coverage ratios of different types of parks (i.e., comprehensive, forest, theme, and historic parks), and the other is the vegetation coverage ratio for all types of green space within the subdistricts where respondents live. The multi-regression results indicate that park buffers (300 m and 500 m) have positive and statistically significant impacts on residents’ self-rated health, while vegetation coverage has a negative impact, ceteris paribus. Different types of parks have heterogeneous impacts on SRH at different distances: comprehensive, forest, and theme parks positively and significantly affect SRH while historic parks make a marginal contribution to SRH. Regression analyses of subgroups shed light on the disproportional health effects of green space among different socioeconomic and demographic groups. Policymakers and planners should be aware of the approach to measuring accessibility, types and features of green spaces, and the heterogeneous socioeconomic attributes of neighborhoods when drafting green space plans to improve public health outcomes.