The emergence of network paradigm has aroused great interest in understanding the network structure related to knowledge flow and regional innovation performance. Based on the perspective of regional heterogeneous innovation network structure, this paper discusses the influence of the structure factors of industry-university-research collaborative innovation knowledge network on regional innovation performance by using the science and technology statistics of 30 provinces in China and the stochastic frontier production function model. At the same time, the influence of regional innovation environment infrastructure on regional innovation performance is analyzed. The empirical results show that there is a positive relationship between network openness and network density and regional innovation performance in network structure, and the effect of network size and network structure hole is not obvious. Regional innovation environment factors also have an important impact on innovation performance. The study found that regional information infrastructure level, regional industrial structure characteristics, regional market openness and other factors have a positive impact on regional innovation performance.