Modeling habitat selection of species with specific foraging requirements is an effective means to evaluate landscape quality for restoration and conservation purposes. Proportional hazards regression, a discrete choice model, was used to develop resource selection functions for breeding Wood Stork in the southern Everglades during the 2006 nesting season. Wood Storks showed the highest probability of habitat selection in a narrow range of shallow water depths. The quadratic form of water depth (depth depth2) was an important indicator of habitat selection with mean water depths between -25 and 25 cm receiving the highest probability of use (a negative water depth is below average ground elevation). Foraging sites within 20 km of nesting colonies were selected over farther sites. Shrub swamp, mangrove swamp and saltwater marsh vegetation types were used in higher proportions than they occurred in the landscape. Results exemplify the importance of shallow water depths near established stork breeding colonies throughout the breeding season.