Abstract:This paper collects relevant data on collaborative innovation efficiency and key elements of 30 provinces and cities in China from 1998 to 2016, using DEA model to measure the regional collaborative innovation efficiency accurately, and using static and dynamic spatial models to conduct empirical analysis on the effect of key elements. Finding: (1) The efficiency of China"s collaborative innovation is heterogeneous, which is shown in Beijing, tianjin and other regions. (2) The key factors have different effects on collaborative innovation efficiency. The influence of industrial agglomeration in the initial stage is not significant positive, the influence in the rapid development stage is significant positive, and the influence in the excessive agglomeration stage is significant negative. The effect of r d intensity was significantly negative. The effect of human capital is significant negative, and the effect of absorptive capacity is significant positive. The impact of infrastructure is significant positive; The short-term impact of the reform depth was significantly negative, while the long-term effect was significantly positive. Foreign direct investment has a significant negative impact; The short-term effect of spillage is significantly positive while the long-term effect is not significantly positive. Based on this, the paper puts forward policy suggestions to effectively promote the efficiency of regional collaborative innovation.