Abstract: Using the big data analysis method of multi-source fusion, such as data mining, theme modeling and network analysis, research was conducted on the policy and interdepartmental collaboration in network space governance, which is of great significance to policy improvement and mechanism optimization. Based on in-depth mining of the network space policy texts of each department, LDA model was adopted to discover and extract low-dimensional subject sets from the texts and analyze policy distribution. Combined with social network analysis（SNA），cooperative network analysis was carried out for the departments involved in the joint publication policy. Finally, six topics were extracted as indicators for evaluating the layout features of cyber space governance policies. This experimental mining of policies and departments in cyber space governance reflects the status quo and demands of the industry, and provides a theoretical basis and decision-making advice for government policy optimization and collaboration.