Abstract: This article describes the spatia1 changes of the economy in Xinjiang at the county leve1 since 1985 by using the ESDA as well as other tools such as Moran's I, Getis-0rd Gi^* and function of variogram and 3-dimension figure. Based on six time discontinuity surfaces, some conclusions are drawn as follows: (1) Considering the overall spatial economic framework, the county level economy of Xinjiang shows a weak trend of spatial self correlation, with cluster function continuously strengthened on the whole. (2) The spatial structure of hotspot distribution tends to have a circular spatial structure centered on Kuitun, Kelamay and Wusu city, and the hotspots have been remarkably increased after 2000, converging in the central part of Northern Xinjiang, Eastern Xinjiang and the northeastern part of Southern Xinjiang. The coldspots converge in the southwest of Southern Xinjiang and tend to extend to the Aksu region in recent years. (3) The development of spatial economic growth framework is likely to be more unstable and heterogenous, and the convergence level is low and tends to be depressed. Hotspot areas are changing frequently without obvious geographica1 concentrations and tend to be converged in the Hami region of Eastern Xinjiang and the Kashi region in the southwest of Southern Xinjiang. The coldspots converge in the Aksu region and the west of the Hotan region in Southern Xinjiang, and the trend of regional difference of economic growth at the county level is opposite. (4) According to the spatio-temporal characteristics, multi-core is a remarkable character in the evolvement of spatial pattern of the economy in Xinjiang. The spatial economic framework is strong and self-organized, but tends to be weakened. The random of the spatial differential patterns keeps low and the mechanism of the structural differentiation caused by self correlation in space is quite remarkable. The spatial difference is low in the SE-NW direction. The heterogeneity of economic development in the NE direction is typical, and the spatial difference is great.

Key words: evolvement of spatia1 pattern, regional economy, spatial association, variogram, Xinjiang