Abstract: Abstract: The implementation of grain direct subsidy policy plays a positive role in increasing grain output, farmers' income, and promoting the development of agriculture. The precision of issuing the subsidy affects its efficiency of supporting agriculture. In order to obtain the accurate winter wheat planting area, improve the efficiency of grain direct subsidy policy implement, and enhance the impact of public funding in supporting agriculture development, a method of evaluating the implementation performance of the policy by using the technology of remote sensing was researched in this study. Suixi County, a main grain production region in the north of Anhui Province, was selected as the study area. GF-1 image with 16-m resolution was used as the data source to extract the winter wheat planting information and exactly calculate the area. There were 2 steps for information extraction and pretreatment: 1) The first was to calculate the area of winter wheat from remote sensing; 2) The second was to get the accurate planting area of winter wheat by interpreting the image and getting the winter wheat planting spatial distribution polygons and the polyline of linear features, analyzing the measured width of linear features, getting the statistical width of the typical linear features, using the buffer analysis method in Geological Information System (GIS), getting the polygons of the linear feature, and then overlaying the two layers and clipping the linear features from winter wheat features. Small features were subtracted by a coefficient in 36 sampling frames to obtain the accurate area of winter wheat. Finally, the adjusted area from remote sensing and the statistically obtained area of grain direct subsidy were compared both in county scale and town scale. Two important conclusions were drawn: 1) In the whole county, the area of winter wheat statistic subsidy was consistent with the adjusted area based on remote sensing. The area of winter wheat statistic subsidy was 1 239.17 km2, and the adjusted area based on remote sensing was 1 227.37 km2. The difference between them was only 11.8 km2, a relative error 0.96%; 2) In the town scale, the absolute of relative error between the statistical and adjusted area were less than 10% in 5 towns, and less than 13% in 8 towns. The Development Zone and Suixi Town had the biggest relative error. This was probably caused by the special types of land use in the two towns. The main types of land use in the two towns was industrial and commercial land use, with small and fragmented distribution of winter wheat fields, leading to difficulty in interpreting winter wheat cultivation area. Totally, the Nash-Sutcliffe coefficient of the statistical and adjusted area among all the towns was 0.90 and the determination coefficient R2 of them was 0.93. In future, the higher spatial resolution and the multi-temporal data can be used to improve the precision of the winter wheat interpreting in special regions, such as the Development Zone and Suixi Town, where extracting winter wheat information is difficult. In practical application, it may be important to combine the basic data such as remote sensing data, the digital products of determining, registering and certifying the rights to manage rural land, with the digital results of spatial distribution of cultivated land together. It will monitor the crop planting condition accurately in house-hold scale, and will help to improve the efficiency of fiscal funds for supporting agriculture.