Abstract: Urban heatisland (UHI) is a typical climate phenomenon, largely attributable to land use and land cover patterns. Land use / land cover conditions in 2010 from remote sensing data were derived by the linear spectral mixture model, to identify correlations between urban landscape patterns and surface temperature all year round by ordinary least squares (OLS) and geographically weighted regression (GWR) models. Temperatures in Shenzhen were found to range from 22.28 to 41.34 ℃, decreasing from northwest to southeast, with significant UHI. GWR model exploring spatialnon-stationary of aregression relationship, performed better than OLS model to simulate quantitative relationships between landscape pattern and surface temperature. Vegetation landscape index could explain spatial distribution of surface temperatures. In this regard, normalized difference vegetation index and leaf area index have the highest correlation rates, of 88% and 90% respectively. Total area, fragmentation, adjacent impervious surface, species, density, spatial distribution, vertical and horizontal growth conditions of dominant vegetation in combination affect the surface temperature. We demonstrate that by avoiding too much concentrated impervious surface reducing neighboring degrees and improving urban green space NDVI and LAI can effectively alleviate UHI.