Summary Connectivity is regarded globally as a guiding principle for conservation planning, but due to difficulties in quantifying connectivity, empirical data remain scarce. Lack of meaningful connectivity metrics is likely leading to inadequate representation of important biological connections in reserve networks. Identifying patterns in landscape connectivity can, theoretically, improve the design of conservation areas. We used a network model to estimate seascape connectivity for coral reef‐associated fishes in a subtropical bay in Australia. The model accounted for two scales of connectivity: (i) within mosaics at a local scale and (ii) among these mosaics at a regional scale. Connections among mosaics were modelled using estimations of post‐larval small and intermediate movement distances represented by home ranges of two fish species. Modelled connectivity patterns were assessed with existing data on fish diversity. For fishes with intermediate home ranges (0–6 km), connectivity [quantified by the index Probability of Connectivity (dPC)] explained 51–60% of species diversity. At smaller home ranges (0–1 km), species diversity was associated closely with intramosaic connectivity quantified by the index dPCintra. Mosaics and their region‐wide connections were ranked for their contribution to overall seascape connectivity and compared against current positions and boundaries of reserves. Our matching shows that only three of the 10 most important mosaics are at least partly encompassed within a reserve, and only a single important regional connection lies within a reserve. Synthesis and applications. Notwithstanding its formal recognition in reserve planning, connectivity is rarely accounted for in practice, mainly because suitable metrics of connectivity are not available in planning phases. Here, we show how a network analysis can be effectively used in conservation planning by identifying biological connectivity inside and outside present reserve networks. Our results demonstrate clearly that connectivity is insufficiently represented within a reserve network. We also provide evidence of key pathways in need of protection to avoid nullifying the benefits of protecting key reefs. The guiding principle of protecting connections among habitats can be achieved more effectively in future, by formally incorporating our findings into the decision framework.