Net primary productivity (NPP) is an important index of ecological environmental change and can be used as an indicator of climate change. Satellite images (moderate resolution imaging spectroradiometer)and ground-based observations of climatology at 17 sites in Great Dun Huang and surrounding areas were jointly used to improve the Carnegie-Ames-Stanford approach (CASA)model and estimate NPP in Great Dun Huang from 2000-2010. We found a clear spatial distribution pattern and discuss relationships between NPP values of oasis savannas and the Gobi Desert. Temporal and spatial patterns in NPP appearsmall blockorstrippedand annual NPP ranges from 0~790.67gCm~(-2)a~(-1) (mean value：55.89gCm~(-2)a~(-1)). Annual precipitation shows a decreasing trend,and ranges from the southeast of the study area at 202.03mm to the northwest at 34.03mm. The annual temperature ranges from the middle presentstrippedwith a high value area of 10.54°C to the two direction of north and south with a low value area of 4.27°C. The study area is mainly influence by annual precipitation,and it is the main control factor in Great Dun Huang at an annual scale. The analysis between NPP with precipitation and temperature shows that correlation with the monthly scale is lower than with the annual scale,indicating that vegetation growth within a month is more sensitive to the joint function of hydrothermal combination rather than either climate factor. Compared with the Gobi Desert,oasis savannas display a time lag and cumulative effects.