In the context of promoting the 15-minute life circle and advocating low-carbon travel, increasing the rate of walking to school through environmental interventions has become an important planning objective. However, existing studies face two main limitations. First, field audits used to measure micro-level street quality are often time-consuming and labor-intensive. Second, insufficient attention has been paid to the relationships between built environment characteristics and walking-to-school rates across various scales. Taking Hong Kong, China, as the study area, this study employs Extreme Gradient Boosting and Multiscale Geographically Weighted Regression models to investigate the relative importance, spatial heterogeneity, and scale effects of built environment and street quality variables. Based on these findings, planning implications are proposed at the macro, meso, and micro levels to support walking-to-school promotion and refined street design.