In the context of increasing fragmentation of global data governance rules, policy cooperation among countries regarding cross-border data flows has evolved into a networked structure. However, the overall structure and formation mechanisms of such cooperation lack systematic quantitative identification. Based on 113 international cross-border data flow agreements from 2015 to 2024, a global cross-border data flow policy cooperation network was constructed. Social network analysis and quadratic assignment procedure regression were employed to examine its structural characteristics, the roles of key nodes, and its formation mechanisms. The results indicate that the overall density of the global cross-border data flow policy cooperation network is low, yet it maintains a single connected component. It exhibits small-world characteristics featuring strong local clustering and high global connectivity," forming a "core-periphery" structure centered on developed economies with the gradual integration of emerging economies. Power differentiation within the network is significant. The European Union occupies a normative core position, while China and the United States primarily lead their respective cooperation clusters. The United Kingdom and Singapore play strategic intermediary roles. The formation of network cooperation relationships follows a dual logic of "similarity and complementarity." Factors such as path dependence, geographical proximity, policy similarity foundational serve as conditions for cooperation by reducing cooperation costs. Differences in digital infrastructure and trade openness serve as core drivers for cross-institutional cooperation. The findings can provide references for identifying key partners and optimizing cooperation pathways in the global competition for digital governance.