The conservation of wide-ranging taxa depends critically on planning efforts that consider both habitat and connectivity needs of focal species.  Fragmented landscapes that include expansive areas of urbanization can further complicate analyses and realistic conservation goals.  Despite these challenges, contemporary efforts tend to rely on overly-simplistic decision rules and tools (e.g., GIS overlays, least-cost pathways, etc).  We believe the use of theoretically grounded spatial tools that permit a more integrated analysis of the landscape are needed in order to produce defensible land-use plans.   We used a suite of habitat and landscape connectivity models to develop a long-term conservation strategy for cougars in a highly fragmented region of southern California.  Within our 35,000-km² study area, we used empirical and expert-based information to derive spatially-explicit models of core and dispersal habitats.  We then integrated these models to predict important linkage zones among core areas using models from electronic circuit theory, which predict movement probabilities given the quality and configuration of dispersal habitat between core areas.  Probabilistic model outputs were used to quantitatively compare the value of alternative pathways, and evaluate the implications of continued habitat loss and fragmentation.  Our results both illustrate an integrated approach to habitat conservation planning, and provide a framework to test a-priori hypotheses regarding animal movement.  The portability of these principles can serve as a framework for long-term planning for this and other species in various regions in North America.