[Purposes]Local climate zones(LCZ)have become the foundation of urban climate research.Understand the patterns of spatial variation in surface temperature within a local climate zone is of great significance for improving the living environment in these areas.[Methods]The study selected three typical local climate zones in Zhengzhou:multi-story residential areas,high-rise residential areas,and institutional land.Using the multi-scale geographical weighted regression(MGWR)method,it explored the relationship between normalized vegetation index(NDVI),green space,shade,impervious surface area,building height,and sky view factor(SVF)with land surface temperature(LST).[Findings]The results showed that the three types of land-use affected land surface temperature were different,and NDVI had the greatest impact on institutional land use areas,with a bandwidth of 27.88 m.The most influential landscape parameter in the high-rise residential is shading with a bandwidth of 2 728.6 m,but the most influential parameter in the multi-storey residential is impervious ground with a bandwidth of 20.93 m.[Conclusions]The influence of 3D landscape and the combination of landscape on urban land surface temperature was clarified,and it was found that the bandwidth of NDVI and building height was less than 30 m,which further narrowed the scale of urban climate research.It shows that the living environment in LCZ can be changed through the adjustment of landscape elements.