Abstract: In order to master the spatial variability of karst rock desertification, and solve the problem of the shadow "missing" information of remote sensing images existing in the karst area，rock desertification indices were extracted in Guohua Ecological Experimental Area，which was in Pingguo county of Guangxi province by applying image processing software of ENVI(environment for visualizing images). And the rock desertification indices of the shadow areas were excluded with the remote sensing images, then their spatial distribution were analyzed with geostatistical method. The results indicated that the rock desertification indices of the study area were impacted intrinsic factors and performed strong spatia1 autocorrelation. The spatial distribution maps of the rock desertification indices interpolated by Kriging interpolation method showed that the higher indices emerged in the northeast and southwest of the study area, and the spatial distribution of the karst rock desertification was apparently strip and block. The spatial distribution of the rock desertification levels in the study area is chiefly controlled by lithology and topography. Using geostatistical methods to predict the shadow "missing" information can provide a new idea and method for monitoring and evaluation of the karst rock desertification.