Food web analysis can be a critical component of ecological risk assessment, yet it has received relatively little attention among risk assessors. Food web data are currently used in modeling bioaccumulation of toxic chemicals and, to a limited extent, in the determination of the ecological significance of risks. Achieving more realism in ecological risk assessments requires new analysis tools and models that incorporate accurate information on key receptors in a food web paradigm. Application of food web analysis in risk assessments demands consideration of: 1) different kinds of food webs; 2) definition of trophic guilds; 3) variation in food webs with habitat, space, and time; and 4) issues for basic sampling design and collection of dietary data. The different kinds of food webs include connectance webs, materials flow webs, and functional (or interaction) webs. These three kinds of webs play different roles throughout various phases of an ecological risk assessment, but risk assessors have failed to distinguish among web types. When modeling food webs, choices must be made regarding the level of complexity for the web, assignment of species to trophic guilds, selection of representative species for guilds, use of average diets, the characterization of variation among individuals or guild members within a web, and the spatial and temporal scales/dynamics of webs. Integrating exposure and effects data in ecological models for risk assessment of toxic chemicals relies on coupling food web analysis with bioaccumulation models (e.g., Gobas-type models for fish and their food webs), wildlife exposure models, dose-response models, and population dynamics models.