Urban green space carbon storage plays a critical role in mitigating the atmospheric CO₂ concentrations and addressing global climate changes. Nevertheless, the disparities in urban green space carbon storage between urban and rural areas, and their causes, remain poorly understood. In this study, using Wuhan's built-up area as a case example, we developed a model to measure urban green space carbon storage and assess its spatial distribution by combining field surveys, remote sensing, and machine learning. Additionally, the direct and indirect separation method was employed to reveal the urban-rural disparity in carbon storage within urban green spaces along the dynamic urban-rural gradients. Furthermore, the underling factors driving the urban-rural differences of urban green space carbon storage were quantified by using correlation and regression analysis. The results showed that: 1) In 2022, the area of urban green space in the built-up areas of Wuhan was 1677 km², accounting for 36.78% of the total study area. Carbon storage stored in the urban green spaces was 10906 Gg (1 Gg=10⁹ g), showing a radial increase from urban cores to the suburbs. The urban green space carbon density ranged from 0 to 383 Mg C/hm², with an average density of 23.92 Mg C/hm². 2) Urban green space carbon storage decreased linearly with increasing urbanization intensity, aligning with the urban-rural gradient. After separating the direct impact of urbanization on urban green space carbon storage, its indirect impact was typically V-shaped (a decrease followed by an increase) with the increase of urbanization intensity. When the urbanization intensity exceeded 0.9, the indirect impacts of urbanization on urban green space carbon storage shifted from negative to positive effects. 3) Urban green space carbon storage was influenced by both the internal characteristics (i.e., landscape structure) and the external environment (i.e., climate change and human activities) of the urban green space system. On the overall urban-rural gradient, Mean Shape Index, Landscape Division Index, Shannon's Diversity Index, temperature, precipitation, and distance from roads were positively correlated with urban green space carbon storage. Conversely, percent of Landscape, land surface temperature, population density, PM_2.5, and night lights were negatively correlated with urban green space carbon storage. Compared with landscape structure and climate change, human activities were the key driving factors driving the urban-rural differences of urban green spaces carbon storage. Specifically impacting factors include land surface temperature and nighttime lights. However, these correlations and impact strengths varied across different levels of urbanization intensity. Overall, our findings provide scientific support for urban decision makers to formulate practical strategies for urban green spaces landscape design and management. This will help achieve urban "carbon neutrality" and sustainable development of future cities.