In this study social network analysis is used to empirically map out existing collaborative social networks influencing water governance in the Mkindo catchment, in Tanzania. I assess how existing social networks may affect the capacity to govern water and explore how knowledge about these networks can be used to facilitate more effective or adaptive water resources management. The study is the first of its kind to apply social network analysis to organizations influencing both blue water (the liquid water in rivers, lakes and aquifers) and green water (the soil moisture used by plants) at the catchment scale. Using a questionnaire and semi-structured interviews, social network data for 70 organizations was generated. With the help of social network analysis software1, the reciprocated collaborative interactions within the water governance network were traced, to better understand how the network facilitates and constrains individual actors, as well as the whole network, in managing water resources. Results show that there is currently no organization that coordinates the various land and water related activities at the catchment scale. Furthermore, there are important actors influencing water governance at the catchment scale, but they are not adequately integrated to the formal water governance system. Water user associations (WUA’s) are in the process of establishment, and could bring together actors presently not part of the formal governance system. However, at the moment the establishment of WUA’s seems to follow a top-down approach that fails to consider the existing informal organization of water users that are revealed through this social network approach. Instead of imposing institutional arrangements I argue that it is more promising and effective to identify and build on existing social structures. Social network analysis can help to identify existing social structures and points for interventions to increase the problem solving capacity of the governance network.