Abstract: In order to improve the cropland productivity in Sanjiang Plain, the spatial distribution of the low-yield cropland in Sanjiang Plain and its influence factors were analyzed with the spatial analysis method based on the MOD17A3 data set of NASA EOS/MODIS(TERRA), DEM(1︰250 000) data, roads, rivers, residents data, and so on. The results showed that the low-yield rate of cropland in the shady slope (32.16%) is higher than that in the sunny slope (25.49%), which showed that sunny slope was beneficial to crop growth. With the increasing of the altitude, the low-yield rate of upland field had a tendency of “low-high-low”. The highest value of low-yield rate was 39.62% and happened in the altitude zonal from 100 to 200 m. The low-yield rate of upland field and paddy field were inversely related to the distance to the residents(≤5 km), roads(≤3 km) and rivers(≤12 km). The correlation coefficients of upland field low-yield rate and the distance to the residents, roads and rivers were -0.979 (p＜0.01), -0.999(p＜0.05) and -0.935 (p＜0.05). And the correlation coefficients of paddy field low-yield rate were -0.980(p＜0.01), -0.998(p＜0.05) and -0.923(p＜0.05). The study can provide scientific basis for ensuring regional agricultural production, reasonable utilization of land resource and regional sustainable development.