Abstract: ： Using statistical and geostatistical methods analyzed spatial variability of surface soil nutrients in cotton planting soil in Wushi county. The research showed：①mean soil pH was 7.8,showing slight alkaline soil feature. Except that the slow potassium content average value was greater than 500 mg/kg,grading was very rich and other elements content grading was to the appropriate or most appropriate. ②There was minimal difference in pH and large difference in available phosphorus content between sampling points;There was a positive correlation and spatial autocorrelation between the elements,but not very close.③organic content was same in the entire sampling area,content level was appropriate;Total carbon content was lower in the north part,other area total carbon content was 0.750~1.000 g/kg,the level of content was appropriate; Alkaline nitrogen content was lacking in the north,suitable in the south; Available phosphorus content was 10.000~20.000 mg/kg,the content level was the most appropriate; Slow potassium content was lower in the eastern part,rich in the northern part,very rich in the central and southwest part. The available potassium content was ranged 74.441~100.000 mg/kg in the southwest and northeast part of the sampling area and content level was suitable,the levels were the most appropriate in northern,southern,southeastern and central part.

Key words: soil nutrients, geostatistics, spatial autocorrelation, spatial variability