Net primary productivity (NPP) is an important indicator of terrestrial ecosystems and accurately estimating NPP is fundamental in terrestrial carbon cycle research. Ningxia Province is dominated by grassland and the estimation of grassland NPP can provide a scientific basis for the sustainable use of grassland resources and the improvement of grassland management. In this study, the grassland ecosystem NPP in Ningxia Province was estimated using a Carnegie-Ames-Stanford Approach (CASA) model and Moderate Resolution Imaging Spectroradiometer (MODIS) data from 2000 to 2015. To study the spatiotemporal characteristics and future evolution trend of grassland NPP, the correlation between the grassland NPP and meteorological factors was analyzed using a simple linear regression model and the Hurst index. The results showed that the CASA model accurately simulated the grassland NPP in Ningxia Province. The correlation coefficient between estimated NPP by the CASA model and measured NPP was 0.93 (P < 0.01), and the spatial characteristics of the estimated NPP image were similar to that of the NPP image of MOD17. In the whole province, the average annual grassland NPP estimated by the CASA model was 148.28 g C m^-2 a^-1 during 2000-2015. Meanwhile, the average annual grassland NPP increased with fluctuations during the past 16 years and the linear growth rate was 3.84 g C m^-2 a^-1 (P < 0.01). Regarding the spatial variation of grassland NPP in Ningxia Province, the NPP in 98% of the grassland increased during the past 16 years and this increasing rate decreased from South to North. The Hurst index of grassland NPP in the past 16 years ranged from 0.27 to 0.81 and the mean value was 0.53, which indicated the grassland NPP will maintain the same trend of change in the future. At an annual scale, the average grassland NPP in Ningxia Province was mainly affected by annual precipitation, and had a weak correlation with the average annual temperature. At a monthly scale, there was also a weak relationship between the grassland NPP and temperature in the growing season; however, the monthly total precipitation exerted an important influence on the grassland NPP accumulation. Furthermore, there was a one-month time lag in the monthly NPP response to mean temperature, but this phenomenon did not exist between the monthly NPP and total precipitation. The main reason for this phenomenon was possibly the hysteresis response to temperature in the steppe and desert grasslands, which were the main types of grassland in Ningxia Province.