Urbanization leads not only to rising temperatures in urban areas but also to dramatic alterations in local microclimatic conditions, which in turn affect human thermal comfort, a crucial factor for health and wellbeing. Although past research has predominantly focused on the correlation between urbanization and ambient temperature, few studies have explored the causal mechanisms of thermal comfort—a more comprehensive and integrated indicator—across varying urban contexts. This gap in the literature is particularly significant as the influence of urban form and land use on thermal comfort is characterized by complex spatial and temporal heterogeneity. Such complexity presents a formidable challenge when attempting to conduct systematic, large-scale, and long-term analyses. To address these challenges, this study adopts innovative analytical methods that incorporate both spatial diversity and temporal dynamics. Specifically, it integrates Multispatial Convergent Cross Mapping (multispatialCCM) and Geographical Convergent Cross Mapping (GCCM) to identify and quantify the nonlinear causal relationships between different types of land use and urban thermal comfort. The multispatialCCM approach utilizes the reconstruction of time series data from various spatial units, which allows for the detection of hidden causal links that standard correlation analyses might miss. Meanwhile, the GCCM method enhances this analytical framework by incorporating geographical adjacency characteristics, thereby increasing the robustness and reliability of the spatial causal inferences drawn from the data. The study focuses on Fujian Province, examining different scales of built-up areas over an extended period from 2005 to 2022, with an additional focus on varied regional contexts in 2022. The analysis reveals that in small-to-medium and medium-built-up areas, there is a significantly positive causal relationship between the expansion of construction land and thermal comfort degradation, with P-values of 0.037 and 0.015 respectively, indicating the statistical significance of these effects. Notably, the adverse impacts of such expansion are even more pronounced in these urban scales. In contrast, within large-scale built-up areas, the presence of bare land is found to exacerbate heat stress, whereas the influence of farmland is not uniform and depends substantially on regional characteristics and local environmental contexts. Moreover, the study highlights that forest cover exhibits the most substantial beneficial effect on improving urban thermal comfort. In comparison, while grasslands, shrubs, and water bodies do contribute to thermal regulation, their effects are generally weaker and subject to variability across different regions; in some cases, these land use types might even have unfavorable outcomes under specific environmental conditions. Based on these findings, the research advocates for urban planning strategies that are tailored to the specific scales of built-up areas. It recommends an increase in forest, grassland, and shrub coverage, alongside an optimized configuration of water bodies, to mitigate the negative impacts of urban heat and to promote more balanced urban microclimates. Overall, this research provides robust scientific evidence for integrating thermal comfort considerations into urban planning and environmental management. The adoption of advanced causal inference methods such as multispatialCCM and GCCM not only deepens our understanding of the complex interplay between land use and thermal comfort but also offers critical theoretical and practical insights for developing sustainable urban environments in the context of ongoing urbanization challenges.