Abstract: This paper uses the non-radial, variable return to scale Super-SBM model to measure the efficiency of China’s industrial green innovation from 2005 to 2019, and analyzes its temporal and spatial differentiation characteristics using global and local Moran indices, and further explores its dynamic transition type and path. The research results show that the overall green innovation efficiency of China's industry is on a wave-shaped upward trend, but there are large differences in time and space. The provinces with high-high (H-H) agglomeration and low-low (L-L) agglomeration are mostly in the eastward direction. The stepped distribution law of the west gradually decreases; the temporal-spatial transition type exhibits a high degree of spatial stability, and the temporal-spatial evolution has a strong path dependence characteristic. The research conclusions of this paper are helpful to clarify the spatial differentiation characteristics and dynamic evolution laws of my country's industrial green innovation efficiency, and then provide useful enlightenment for the promotion of regional innovation drive and the implementation of green development strategies.