In the study of spillover effect of urban innovation space, spatial measurement is sensitive to the choice of spatial weight. Based on the panel data of 271 major cities in China from 2003 to 2018, this paper proposes an improved spatial weight by designing an indicator of administrative regulation, and compares the spatial autocorrelation of innovation by introducing strong and weak administrative adjustment factors to construct improved spatial weights. Furthermore, the influence of factors on urban innovation is empirically examined by using the Durbin model and decomposing correlation of urban innovation spatial in order to explore the different effects of administrative regulation. The results show that there are spatial dependence and mutual promotion of urban innovation in China during the investigation period, and the spatial spillover of urban innovation is also affected by the administrative boundary effect. In the geographical distance space, urban population can promote innovation spillover among cities, and administrative regulation will weaken this spillover effect. Industrial specialization and diversification agglomeration inhibit innovation spillover, which will be significantly weakened by administrative intervention. In the economic distance space, industrial specialization agglomeration and other factors can significantly promote innovation spillover between cities, while administrative regulation has little effect. The spatial spillover differences of various influencing factors provide theoretical and empirical basis for governments at all levels to formulate innovation policies and coordinate practice.