Net primary productivity (NPP),one of the global change research hot spot,can represent the production capacity of vegetation. Taking HulunBuir as the study area,MODIS-NDVI,landcover dataset and meteorological data were used to estimate NPP based on an improved Carnegie-Ames-Stanford Approach model (CASA),and to analyze the temporal-spatial characteristics of vegetation NPP in HulunBuir from 2000 to 2015. The results showed that:(1) The spatial pattern of NPP in HulunBuir showed a good strip regularity that there was an obvious increase trend from southwest to northeast. (2)The range of NPP in HulunBuir was mainly between 200 and 400gC /(m~2·a) with an average annual NPP of 340.7gC /(m~2·a) from 2000 to 2015. Although not too much high values of NPP were found in HulunBuir,its proportion was increasing year by year. (3)From 2000 to 2015,the NPP of HulunBuir showed an decreasing trend in the first few years while an increasing trend in the rest years. (4)For different vegetation types,the NPP of cultivated land showed an continuous increasing trend while native grassland showed an continuous decreasing trend from 2000 to 2015. The NPP of other vegetation types showed an increasing trend on the whole with fluctuations in some years.