---

Land surface temperature is an important indicator in the study of ground energy balance models on a regional and global scale. Due to the limitations of meteorological stations, remote sensing can be a good alternative to estimating surface temperature. In the present study, the relationship between different land uses and ground surface temperature in Kosar city and spatial autocorrelation has been investigated using Moran index. Landsat 8 (OLI) and Landsat 5 (TM) satellite images were used for this purpose. The present study was carried out in four main stages: a) preparing images for 1987 and 2015, b) making the necessary predictions, c) preparing a classification map using the object-oriented method, and d) extracting the earth's surface temperature with Separate window algorithm. The results showed that dryland and rangeland agricultural uses in 1987 and 2015 have higher temperatures than other uses, due to the dryness of selected agricultural products in the selected period (June), which caused the temperature to rise. Is. In contrast, the use of water areas due to their high heat capacity in these years have the minimum average temperature. In addition, hot and cold clusters of Kosar thermal islands were extracted using the hot spot analysis index (Hotspot). Finally, the analysis of spatial autocorrelation with global Moran indices showed that the surface temperature of Kosar city has a spatial structure and is distributed in the form of clusters.