In disaster response, collaboration facilitates interactions among actors, such as the government, the military, nongovernmental organizations, and civil society organizations. This study examined the longitudinal changes in collaborative governance in Myanmar’s disaster responses based on cases of flooding in 2015, 2016, and 2018. To examine the mechanisms underlying this dynamic network formation, the collaborative ties of the actors involved in search and rescue activities were converted into longitudinal relational data sets, and the evolution of collaborative governance was analyzed by relying on the assumptions of social capital, transaction cost, homophily, and resource dependency theories and using a longitudinal social network analysis method. The findings show that the collaborative networks of search and rescue processes in disaster response evolved and changed over time according to the hypothesized patterns of strong, weak, and preferential tie formations. The study also revealed that the collaborative governance system assumes the form of a hierarchy rather than a generalized exchange, and the actors’ reliance on military organizations is not obvious due to the emerging alternative non-military actors and diverse local actors observed in the cases.