Abstract: Desertification is the prominent eco-environmental problem in semi-arid and sandy land area. And land use is the main driving force of desertification. The development process of desertification has been primarily curbed in general but kept expanding in partial locations. At the circumstance, it is significantly beneficial for anti-desertification and ecological restoration to study the cultivated land use characteristics based on cultivated land particles in the semi-arid and sandy land area. Materials applied in this paper included Landsat-8 remote sensing images, land use/cover maps, atmospheric water vapour content, DEM and soil nutrient data. Split window method was used to calculate land surface temperature of cultivated land. And characteristics of land surface temperature of cultivated land and its causes were analyzed. The results showed that the range of land surface temperature of cultivated land was wide. The minimum and maximum of land surface temperature of cultivated land in row/path 030/120 remote sensing image were 294.03 K and 308.61 K. And those in row/path 030/119 remote sensing image were 291.43 K and 310.62 K. The temperature ranges whose cultivated land area accounting for more than 10% of the total cultivated land, included 297 K, 298 K, 298 K, 299 K, 299 K, 300 K, 300 K, 301 K, and 301 K, 302 K in row/path 030/120 remote sensing image. And those were 300 K, 301 K), 301 K, 302 K), 302 K, 303 K), 303 K, 304 K) in row/path 030/119 remote sensing image. The spatial variation of precipitation resulted in a large spatial variation of crop growth condition in Horqin Left Back Banner, which resulted in a large range of land surface temperature of cultivated land. As to the temperature difference of cultivated land particles, land surface temperature range of cultivated land particles in row/path 030/120 remote sensing image was between 0 and 10 K and that in row/path 030/119 remote sensing image was between 0 and 11 K. This indicated that land surface temperature difference of cultivated land particle was large. Furthermore, the area of cultivated land particles with large land surface temperature difference took a large proportion. The areas of cultivated land particles with land surface temperature difference greater than 3 K in row/path 030/120 and 030/119 remote sensing images took percentages of 36.895% and 57.971% respectively. The areas of cultivated land particles with land surface temperature difference greater than 4 K in row/path 030/120 and 030/119 remote sensing images took percentages of 16.841% and 41.571% respectively. This was attributed to the redistribution of precipitation along dune-interdune and the elevation difference of cultivated land particle. As to the land surface temperature at cultivated land particle margin, among 22107 cultivated land particles in row/path 030/120 remote sensing image, average land surface temperature at the marginal area of cultivated land was greater than that of the cultivated land particle in 13268 cultivated land particles. And among 1644 cultivated land particles in row/path 030/119 remote sensing image, average land surface temperature at the marginal area of cultivated land was greater than that of the cultivated land particle in 866 cultivated land particles. Both areas took percentages of 73.71% and 53.80% respectively in both cultivated land areas. And micro-topography, soil quality and management measures induced this phenomenon. According to the results above, we reckon that grid method should be adopted in cultivated land quality evaluation in semi-arid and sandy land area. Cultivated land at the margin of cultivated land particle should be de-farmed in order to anti-desertification and ecological restoration.