AbstractSpawning habitat, the cornerstone of self‐sustaining, naturally reproducing walleye Sander vitreus populations, has received limited quantitative research. Our goal was to quantitatively describe the structure and quantity of natural walleye spawning habitat and evaluate potential selection of habitat in Big Crooked Lake, Wisconsin. In 2004 and 2005, we located and delineated walleye egg deposition polygons through visual snorkel and scuba surveys. We also delineated recently deposited, adhesive egg patches daily along one spawning reef in 2005. To determine habitat selection, we quantified and compared spawning and lakewide available habitat at different scales. In both years, walleyes used similar spawning habitat, including three geomorphic types: linear shorelines, a point bar, and an island. Walleyes used only 14% of the entire lake shoreline and 39% of the shoreline comprised of gravel (6.4–76.0 mm), cobble (76.1–149.9 mm), or coarser substrates for spawning in 2005, indicating selection of specific spawning habitat. Lakewide, walleyes spawned close to shore (outer egg deposition polygon boundary mean distance = 2.7 m), in shallow water (outer egg deposition polygon boundary mean depth = 0.3 m), and over gravel substrate (percent coverage mean = 64.3) having low embeddedness (mean = 1.30). Our best nearshore (0–13‐m) resource selection function predicted an increase in the relative probability of egg deposition with the increasing abundance of gravel, cobble, and rubble (150.0–303.9‐mm) substrates and a decrease with increasing distance from shore and water depth (89.9% overall correct classification). Adhesive egg patches confirmed that walleyes actively chose nearshore, shallow‐water, and coarse‐substrate spawning habitat. The quantitative habitat information and predictive models will assist biologists in developing walleye spawning reef protection strategies and potentially aid in designing and evaluating artificial spawning reefs.