Taking the cut-off years of China’s four five-year development strategies as the research time point, based on the radial function (RBF) neural network model, China’s 337 municipal administrative units (except Hong Kong, Macao and Taiwan) construction land utilization strength in 2000, 2005, 2010 and 2015 was measured, and the exploratory spatial data analysis model (ESDA) was used to analyze its spatial and temporal evolution pattern. The main driving factors of spatiotemporal differentiation of construction land use intensity in China were analyzed by using geographically weighted regression model (GWR). The results show that from 2000 to 2015, China’s construction land use intensity showed an overall upward trend, and the spatial difference was significant. The high-intensity utilization level was mainly concentrated in the Bohai Rim Economic Zone, major coastal cities and provincial capital cities. The intensity of china’s construction land utilization presented spatial self-correlation, and the global pattern of separation was basically consistent with the pattern of cold hot spot distribution and spatial heterogeneity. The spatial pattern evolution of China's construction land use intensity was mainly affected by four factors: economic development level, industrial structure, urbanization level and government economic regulation. The rank of its influence degree was urbanization level < industrial structure < level of economic development < government economic regulation.

以中国4个五年发展战略的截止年份为时间节点, 基于径向函数(RBF)神经网络模型测度2000, 2005, 2010和2015年中国337个市级行政单元(不含港澳台地区)建设用地的利用强度, 并利用探索性空间数据分析模型(ESDA)分析其时空演变格局, 运用地理加权回归模型(GWR)分析中国建设用地利用强度时空分异的主要驱动因素。结果表明: 1) 2000—2015年, 中国建设用地利用强度整体上呈上升趋势, 且空间差异显著, 高强度的利用水平集中在环渤海经济区、各大沿海城市和省会城市; 2) 中国建设用地利用强度呈现空间自相关性, 且全局分异格局与冷热点分布格局及空间异质性的态势基本上吻合; 3) 中国建设用地利用强度的空间格局演变主要受经济发展水平、产业结构、城市化水平以及政府经济调控4 种因素影响, 影响程度的排序为城市化水平<产业结构<经济发展水平<政府经济调控。