Metapopulation dynamics depend on the exchange of individuals between populations across the landscape. The environment that the migrants must traverse is influenced by many forces, so the connections are often complicated pathways, which can be represented as a network. The structure of these networks will determine which populations will receive more migrants than other populations, and this in turn affects the lifetime of the metapopulation. We present a modification of the Drechsler (2009) formulae for the mean lifetime of a metapopulation based on network properties where the arrangement of the population connections is not limited to simple dispersal kernels. We provide a graph-theoretical framework for analysing the dispersal network and apply this model to the small-world network of the Great Barrier Reef as well as to conservation planning in general. Our results highlight that the topology of the dispersal network strongly influences the metapopulation mean lifetime. Metapopulation models and conservation planning need to include the capacity for basing interactions on complex topological structures. Copyright © 2010 Inter-Research.