The oceans provide key ecosystem services for millions of people around the world and it is important to gauge how these services will change twenty, fifty, and hundred years from now. Any reasonable answer to this question must account for both the complexity and adaptive nature of marine social-ecological systems. Here I will touch upon a number of studies that focus on the complexity and adaptive nature of marine systems. First, a spatially explicit nearshore metapopulation model is used to identify subpopulations that are good candidates for marine reserves. Second, a coral metacommunity model is used to identify the role that dispersal plays in determining alternative stable states, and a spatial implementation of the model specific to the Coral Triangle is used to identify the sequence by which subpopulations are likely to flip to algal-dominated states under climate change. Third, the social-foraging tactics that fishermen employ to maximize their harvest — whether cooperative or not — is investigated using idealized agent-based simulations and a spatially implicit model. These results aid in the implementation of evolutionary game-theoretic simulations, used to identify the economic conditions for stable fishery cooperatives. Fourth, the dynamics of weather insurance cooperatives are explored using Myanmar aquaculture communities as a case-study. Last, early work from a new project focusing on using Manifold Learning to develop new early-warning signals of critical transitions in multi-scale systems is presented. All these studies address key dimensions of complexity and adaptation in marine systems and I end with a discussion of their relevance to the resilience of coastal communities.