In order to achieve the carbon reduction target and carbon neutrality vision, based on panel data from 30 provinces and cities in China, building the semi-parametric space panel vector autoregression model (SSPVAR), using the impulse response function and derivative scatterplot empirical analysis on the bidirectional spatial conduction effect among technological innovation, the rationalization of industrial structure and advanced industrial structure, carbon emissions, and the nonlinear effects of environmental regulation. The results show that :(1) There are time lag effects and spatial transfer effects among technological innovation, industrial structure upgrading and carbon emissions.Technological innovation and carbon emission have significant positive self-reinforcing effect and positive spatial spillover effect, while rationalization of industrial structure and upgrading of industrial structure have negative spatial spillover effect. (2) Technological innovation promotes the rationalization and upgrading of local and neighboring industrial structure, but both the rationalization and upgrading of industrial structure have different degrees of time lag inhibition on local and neighboring technological innovation. (3) Technological innovation reduces local carbon dioxide emissions, but is not conducive to neighboring carbon reduction; The rationalization of industrial structure and the upgrading of industrial structure are beneficial to the carbon emission reduction of neighboring areas, but increase the local carbon dioxide emission, and the upgrading of industrial structure has a more significant carbon emission reduction effect than the rationalization of industrial structure. (4) Environmental regulation has a significant nonlinear impact on technological innovation, industrial structure upgrading and carbon emissions.