The spatial variation of precipitation and erosion sediment production in loess areas of Wuding River basin in the middle reach of the Yellow River were studied by using Thiessen polygon weighted coefficient of variation method, and the correlation between the spatial variability was explored by using linear regression analysis. Based on Wuding River basin, one of the main sediment source area of the Yellow River, this paper analyzed the temporal and spatial variation of precipitation and sediment yield of the loess area in 1959-2015 before and after the soil and water conservation measures, choosing flood season precipitation (P_flood) and flood season rainfall erosivity (R_flood) for the precipitation characteristics, and specific sediment yield (SSY) for sediments production. The results showed that:(1) Before soil and water conservation measures (1959-1970), the trend of precipitation characteristics in this region was consistent with the change of SSY in time, and the two precipitation characteristics showed a significant power function correlation (P<0.01). The effects of P_flood and R_flood on erosion sediment yield were consistent. After large-scale soil and water conservation measures (1971-2015), due to the construction of a large number of silt-dam in study area, sediment yield decreased sharply, and there was no significant correlation between precipitation characteristics and SSY. (2) On the multi-year time scale (1959-2015), the spatial variability of P_flood was 8%, and the spatial variation coefficient of R_flood was 15%, and the spatial variability of P_flood was less than that of R_flood. The spatial variability of SSY between 1959-1970 was smaller than that between 1971-2015. During 1959-1970, the spatial variability of precipitation characteristics and SSY showed a significant quadratic polynomial correlation (P<0.01), and the correlation between R_flood and SSY was more significant (R²=0.76, P<0.01). There was no significant correlation between precipitation characteristics and SSY from 1971 to 2015. The results showed that the spatial variability of SSY before human activities was mainly affected by the spatial variability of R_flood, while the spatial variability of SSY as mainly affected by human activities after soil and water conservation measures.