Focusing on the inequity of heat distribution and the influence of spatial patterns of urban green infrastructure (UGI) on this distribution, this study used bootstrapping bivariate Pearson's correlation to examine the relationship between sociodemographic variables at the census tract level and land surface temperature. The study assessed the performance of multiple models and utilized the Gradient Boosted Decision Trees model to investigate the effects of UGI spatial patterns on thermal environments. Additionally, this study explored how the spatial patterns of UGI differ among sociodemographic groups and their role in heat inequality. The findings revealed consistent heat disparities among certain population groups. The UGI area and patch density exhibited generalized nonlinear effects on urban thermal environments, whereas more fragmented distribution of UGI patches, along with inadequate provision, implicated higher temperatures in marginalized communities. Therefore, it is crucial to adopt a context-specific strategy to address urban heat inequity effectively. Improving the spatial patterns of UGI and increasing UGI coverage in marginalized communities can help mitigate heat distribution inequities.