Taking the authorized number of invention patents of China's 29 provinces during 2000 to 2015 years as the sample data, this paper carries out stationarity test and Grainger causality test for sample data by using Eviews software. On this basis, it constructs regional innovation linkage network. This paper uses social network analysis method and Ucinet software to analyze the development of regional innovation linkage network in China.The empirical study shows that the overall density of regional innovation correlation network in China is 0.1687, and the density of each subgroup is still quite different. The density of the first, fourth subgroups is obviously different from the density of the second, third subgroups, and the number of the fourth subgroups and provinces is the most. The innovation related relationship among provinces in China's regional innovation linkage network is uneven, and the fusion degree between subgroups is also poor.