Abstract:In this paper,DEA-BCC model and DEA-Malmquist index model were used to calculate the operation efficiency of crowd innovation space in 31 provinces and cities in China from 2016 to 2017. The results show that the overall operation efficiency of mass innovation space in China is good, and the main reason for the low efficiency of mass innovation space in some regions is the low scale efficiency. Further, the national crowd innovation space is divided into four types by cluster analysis: relatively high-efficiency type, scale low-efficiency type, technology low-efficiency type and complete low-efficiency type, and the input and output of crowd innovation space are adjusted by projection analysis. Finally, through qualitative comparative analysis of fuzzy sets (fsQCA), two types of antecedent configurations that trigger the high operating efficiency of mass innovation space are proposed, and suggestions for further improving the operating efficiency of mass innovation space are put forward.