Abstract: Lakes are valuable water systems,used for production of drinking water,fisheries and recreation,and can be an important indirect indicator of climatic change MODIS data is suitable for regional to global operational observation as moderate resolution imaging spectroradiometer,costing little and having high temporal coverage.But the mixed pixel is a common problem.The objective of the study is to develop high accuracy sub pixel mapping algorithm.Firstly,we use the multi endmenber linear spertral unmixing technique to estimate the lake fraction in the each mixed pixel,and develop an algorithm,to locate spatially the water body within mixed pixels corresponding to the water proportion in the pixed.The algorithm produces a finer sub-pixel lake map relative to the source MODIS image.The results indicate that it is an efficient and promising sub pixel map ping algorithm.