The Urban Heat Island (UHI) effect presents a critical challenge to urban sustainability and public health, prompting a search for effective morphology-based mitigation strategies. While urban morphology and vegetation coverage are known to modulate Land Surface Temperature (LST), few studies have comprehensively analyzed their nonlinear interactions at a macro-scale across diverse climate zones. To address this gap, we propose an interpretable machine learning framework that outperforms conventional regression in handling urban complexity. This study quantifies the contextualized influence of urban morphology (2D/3D built-form metrics) and vegetated coverage Enhanced Vegetation Index (EVI) on LST across the U.S. mainland. Results reveal that while Average Building Height (ABH) positively correlates with LST, vegetation coverage generally provides a cooling effect. Crucially, these impacts exhibit distinct climatic thresholds: in Arid zones, effective cooling is limited to an EVI range of 0-0.52, while Cold Temperate zones show diminishing marginal utility, peaking at an EVI of 0.65. Spatial partial derivative analysis further indicates that increasing vegetation in low-EVI regions yields the most rapid LST reduction, particularly in Warm Temperate zones. In contrast, the warming impact of Building Area (BA) is most pronounced in regions with already high density. These findings highlight the limitations of uniform strategies and emphasize the necessity of climate-sensitive planning to balance development intensity with ecological sustainability.