Abstract: Taking the waterfront space around Dianchi Lake in Kunming as the research area, based on POI big data, the nearest neighbor distance, kernel density and standard deviation ellipse of ArcGIS spatial analysis were used to characterize the differentiation pattern of recreation utilization. The study found that the recreational use of the waterfront space around Dianchi Lake has the characteristics of small aggregation and large dispersion, and there were significant differences in the number of various recreational resources. The nearest neighbor ratio of all kinds of recreational use and overall recreational use was less than 1, and there was obvious agglomeration in spatial distribution. In Nanping Street, there was a 'circle-layer' pattern, and different types of recreation use were significantly different in structure. All kinds of recreation use were 'northwest-southeast' deviation, and the agglomeration trend and scale were obviously different.