Restoration of the world’s rivers has become a billion-dollar-per-year global industry, with a common objective being the recovery of endangered or economically important species, typically fish. Predicting how target species respond to restoration has traditionally focused on the direct effects of restoration on these organisms; however, restoration also influences the larger ecosystem or food web within which these species participate, which, in turn, can influence focal species via a variety of indirect pathways. Predicting how a particular species may respond to restoration therefore requires the development of holistic approaches that account for these broader food-web interactions. Forest Service scientists constructed a model and explicitly linked food-web dynamics to environmental factors known to influence stream food webs. Their simulations illustrate that food-web models can be used to explore responses to a variety of river restoration actions, from those that represent relatively direct manipulations of the food web (for example, salmon carcass addition), to those that focus on modifying the physical template of the river upon which these webs of interaction emerge (for example, floodplain reconnection). Model simulation results emphasize that restoration actions can influence stream ecosystems through multiple pathways, and that responses to restoration are context dependent. In the Methow River in northcentral Washington, for example, some river segments are more sensitive to floodplain reconnection, whereas others are more sensitive to riparian restoration or nutrient augmentation via salmon carcass addition. These results illustrate that different locations in the river network will not respond equally to the same restoration strategy and that local environmental conditions and limiting factors should be considered when selecting which strategy or strategies are most appropriate. Thus, this modeling tool can be used to discern appropriate restoration actions among reaches within a watershed, and should be used as one component, among a suite of tools, for regional decisionmaking.