<p>Urban green spaces are vital for promoting residents’ well-being, yet enduring disparities in their spatial distribution remain a significant challenge. Existing studies on green space accessibility models have frequently neglected two critical aspects: the quality of green space and the heterogeneity of urban transportation networks. To address these gaps, this study examined the central urban areas of Chongqing—a mountainous city in China—by combining machine learning-driven street-view recognition with real-time analysis via an Application Programming Interface. Using an indicator-based framework, we evaluated and compared accessibility and equity of green spaces in terms of both quality (quality-adjusted) and quantity (area-based) across three transportation modes: walking, cycling, and driving. Random forest and Pearson correlation analyses were applied to identify the determinants of these the equity patterns. Results indicated that (1) The inequity reduction was more substantial in qualitative accessibility than in quantitative accessibility. (2) Green space accessibility and equity in central Chongqing were lowest for cycling, intermediate for driving, and highest for walking. (3) Housing prices and the proportion of elderly residents were the primary indicators influencing accessibility and equity across walking, cycling, and driving accessibility inequity, while terrain slope significantly exacerbated cycling accessibility inequity. This study advances indicator-based methodological frameworks for assessing green space accessibility, providing actionable insights for urban planners and researchers on addressing terrain-related challenges and socioeconomic disparities within mountainous cities. These findings underscored the necessity of prioritizing quality improvements and mitigating inequities in urban green space planning.</p>