In recent years, Chinese crime information has become more transparent and open than ever, thus providing an excellent opportunity for urban crime study by academics. To obtain a better understanding of the spatial pattern of urban crime, the city of Beijing is chosen as a study area and GIS software is employed to collect spatial data. The authors try to establish the quantitative representation of geographical characteristics of crime associated with urban space in order to create a geographical model of urban crime and space. The authors find that the overall spatial distribution of urban crime in Beijing displays a picture of polycentric structure and distance decay, and that the spatial distribution of urban crime has a reference to traffic centres, concentration of urban commerce and population migration. The numbers of suspects and locations where different types of crime happen have inter-annual variation, while both the total number of crime sites and their criminal density within each district are relatively stable. The authors point out that the spatio-temporal characteristics of sites act on both the participants in property crimes and the criminal factors, which will decide whether or not an offender can commit a crime successfully. The hot spots and periods of time of urban crime in Beijing have a close relation to the fact that success in committing a property crime is based on certain conditions of sites and times which appear in Beijing’s environment. In this sense, socio-spatial dialectic provides a better understanding of the dynamics of China’s crime space. Finally, the spatial anti-crime strategies for urban crime and the insufficiency of research are discussed.