Abstract: The thermal effect of landscape patterns from 50 villages in northern plain of Shaoxing County, Zhejiang Province was analyzed based on 2002 Quickbird and ETM images data. In the core area (build-up area) of villages, 45 landscape patterns indices (LPI) of 5 patches (land-use types including building, water, grass plot, woodland and green area mixed by grassplot, woodland) and whole landscape were calculated, and their correlation to environment brightness temperature (EBT) were discussed. The result showed that there was a sound linear relationship between the LPI and BET; the LPI studied here could be classified into 6 groups, i.e. density index, coverage index, shape index, core area size, indices of 100 m wide buffer ring and others, of which the indices type of density, coverage aggregation of dominant patch or of whole landscape have more significant correlation with BET; the indices about building patch and core area size (coverage, mean patch size and aggregation of building patches, area, perimeter and mean patch size of core area) and the shape indices of temperature-reducing patches had positive correlation with BET, others had negative correlation. So those villages, small size in built-up area, small size and even disperse in patches and narrow shape in buildings, were helpful for environment temperature reducing. 4 empirical models were established (R2>0.9) referring to 4 groups of LPI for predicting and adjusting the EBT being selected. However, the models fitted with single landscape quantity structure index had low performance. It means that the contribution of landscape patterns indices to environment temperature, up to about 25% of total, can not be ignored in village built-up area.