Understanding the impact of land-use change on Net Primary Productivity (NPP) is important in the elucidation of the relationship between the carbon assimilation ability of vegetation and land-use changes at the regional scale, which is also critical to the maintenance of regional ecological safety. Using Beijing as a case study, we analyzed the spatiotemporal distribution of NPP, and the impact of land-use changes based on land-use maps and MODIS data downscaled by 30-m resolution Landsat TM images from 2000 to 2012. The results revealed significant linear relationships between Landsat NDVI and MODIS NPP (R² values between 0.22 and 0.68, P < 0.01) and these linear functions were used to downscale the NPP data for this region. The NPP levels were also significantly affected by the area allocation and distribution of different land-use types. From 2000 to 2012, areas of woodland and grassland increased rapidly, whereas the area of farmland declined. Overall, the NPP of Beijing increased from 1.23 million t C to 1.9 million t C, and 66% of NPP was contributed by woodlands, followed by farmland (26%) and grassland (8%). For 2000 to 2006, land area variations benefited NPP growth and contributed 34% to its change, whereas the variations from 2006 to 2012 impeded NPP growth and contributed 27% to the change. Moreover, reduction of patch aggregation, and increment in patch density and patch richness density contributed to the enhancement of NPP. The results suggest that maintaining the current vegetation quantity and quality, and increasing local landscape heterogeneity and diversity would improve NPP in Beijing.