Using remote sensing data and ground meteorological station data, sand land net primary productivity (NPP) during the growing season from 2000 to 2013 in Hunshandak Sandyland were estimated using CASA model, and the spatial-temporal characteristics of NPP were analyzed. Impacts of climatic factors and human activities on the sand land NPP were primarily revealed by statistical analysis. The results showed that the sand land NPP had increased by fluctuation during growing season in Hunshandak Sandyland from 2000 to 2013,with an average annual NPP of 239.8 gC · m~(-2) · a~(-1) in the study area. The area of high NPP (NPP > 150 gC · m~(-2) · a~(-1)) had increased while that of low NPP (NPP <150 gC · m~(-2) · a~(-1)) had decreased. In spatial pattern, the NPP in northern, central and southern fringe regions of the study area had increased, but no obvious trends had been observed in eastern and some of the western parts. As a whole, the annual NPP had depended more frequently on precipitation fluctuation. The correlation coefficient between NPP and precipitation was 0.86,indicating precipitation was the dominant driving force in vegetation dynamics. The correlation coefficient between NPP and temperature was -0.42. While temperature, sown area of feeding, the total livestock at the year-end and the number of sheep in Hunshandak Sandyland showed a decrease trend,the sand land NPP had increased rapidly and remarkably.