<p>Parks are a crucial element of urban landscapes. Understanding the role of the built environment in terms of park accessibility is vital for promoting travel equity and shaping public policies to improve access to urban parks. This study employs a geospatial artificial intelligence (GeoAI) approach to examine the impact of the built environment on park accessibility at both macro and micro levels. We find that accessibility to urban parks varies significantly across different travel modes, with walking being the least equitable due to its most clustered distribution pattern. While the macro-level built environment provides substantial explanatory power for accessibility across travel modes, our GeoAI-based analysis offers a more nuanced understanding of streetscape qualities, by closely reflecting a more realistic human perception of the built environment. These qualities - street network interconnectivity, safety, openness, spaciousness, greenery, and distinctive signage - significantly correlate with enhanced park accessibility across all travel modes, thereby complementing macro-level factors. Ultimately, this study suggests that placing emphasis on multimodal travel can enhance travel equity, thus providing the relevant authorities and policymakers with cost-effective strategies for improving park accessibility.</p>