The purpose of this study is to data mining for urbanization, evaluate the matching of urban population - land in China from the perspective of temporal sequence and spatial pattern, and put forward some suggestions on optimization of population and land area. By constructing the regional spatial matching evaluation model and regional comprehensive mismatch degree evaluation model, use the 2005-2014 China provincial urban population - land panel data, and evaluate the matching types and grades of urban population - land in China from the provincial scales, then using GIS technology to analysis the provincial urban population - land spatial characteristics. Results show that urban population - land comprehensive in Eastern China shows a fluctuating state, and central and western China are relatively stable since 2005; urban population - land comprehensive spatial matching degree of eastern China is the highest, central China follows, western China is the lowest, according to matching level spatial distribution characteristics, relative matching type of provincial accounted for a relatively small proportion, generally concentrated in eastern and central China, proportion of low mismatch type provincial is the largest, and basic maintenance is around 17 provinces, high mismatch type provincial proportion is the smallest, usually no more than two and located in the western region. Therefore, facing spatial matching problem of urban population - land in China, we should gradually build urban population - land spatial matching evaluation mechanism, reasonably evaluate types and grades of regional urban population - land spatial matching, formulate differentiated policy for population - land management, optimize spatial allocation of urban population - land, and focus on connotation of urban population - land spatial equilibrium, to build a harmonious relationship between people and land.