Abstract: Using panel data from 30 provinces in China over the period 1995-2015, introducing the time factor into the spatial panel model, a dynamic spatial Dubin model with a different spatial weight matrix was built based on the bias corrected QML estimator method to examine the spatial spillover effects of infrastructure and human capital on regional innovation capability. The results show:(1)the traffic infrastructure and human capital in different spatial weight matrixes have significant spillover effects on regional innovation ability:the transportation infrastructure has a diffusion effect, while human capital an agglomeration effect;(2)the spatial spillover effects of infrastructure and human capital on regional innovation capability are subject to geographical distance and economic development level;(3)the effects of traffic infrastructure and human capital on regional innovation capability are variable;(4)the traffic infrastructure and human capital have no significant impact on regional innovation speed.