Wetlands are rich in biological diversity and high productivity of the ecosystem,they play an important role in ecosystem carbon cycle.Wetland vegetation net primary productivity(NPP)is an important indicator of the health of wetland ecosystems.In this paper,combined with field surveys,laboratory measurement and remote sensing technology,the three kinds of typical wetland vegetation——〖WTBX〗Spartina alterniflora,Phragmites australis〖WT〗 and 〖WTBX〗Scirpus mariqueter〖WT〗 NPP were studied in Shanghai Chongming Dongtan.Firstly,selected three kinds of typical wetland vegetation type,the biological characteristics parameters of the vegetation was surveyed,including fresh weight of vegetation,height,density,coverage and leaf area index(LAI),etc..The regression model between NPP and LAI was set up in the sample of different vegetation type.Secondly,using SPOT5 images,after geometric correction and radiometric calibration,the image was classified by objectoriented,and modified adjust soil vegetation index(MSAVI)was calculated,which can reflect the elimination of soil and vegetation characteristics of the background factors.Then estimation LAI model was set up with SPOT5 MSAVI.Finally,according the NPP regression model and LAI estimation model,different wetland vegetation NPP model was set up at regional scale in Chongming Dongtan.It provides a good method and theoretical base for typical wetland vegetation structure and function.