Walleye Sander vitreus are commonly introduced to reservoirs and provide valued fisheries. Walleye are lost though harvest and natural mortality in all systems, but escapement may also reduce reservoir populations. However, the effects of harvest, natural mortality, and escapement loss on reservoir Walleye populations are not known. Our objective was to quantify Walleye harvest, natural mortality, and escapement in two Iowa reservoirs to understand how these factors limit populations. We also assessed Walleye behaviors and compared them among escaped, harvested, and in‐lake Walleye to determine if behavior was associated with fate. Walleye were radio tagged beginning in October 2016 and tracked through May 2019. Telemetry was used to estimate annual core and home ranges, seasonal depth use, and movement rates. Detection histories were analyzed using a multi‐state capture‐recapture model with environmental covariates to estimate weekly harvest, natural mortality, and escapement. Harvest was greater in Big Creek than Brushy Creek and was higher in April through July compared to other months. Natural mortality was positively related to water temperature but did not differ between systems. No Walleye escaped from Big Creek (fish barrier on spillway) but escapement at Brushy Creek (no fish barrier) was positively related to mean water levels during April. During 2017‐2018, annual Walleye escapement ranged from 22‐47% at Brushy Creek (0% at Big Creek), harvest ranged from 13‐27%, and natural mortality ranged from 36‐38%. Walleye depth was shallower during spring and summer when escapement occurred, intermediate in fall, and deepest during winter. Escaped Walleye at Brushy Creek used greater depths and moved less than in‐lake Walleye; however, other behaviors did not differ among Walleye groups. Our results suggest escapement can have a larger effect than harvest on reservoir Walleye populations in systems without physical barriers and should be taken into consideration when managing reservoir Walleye fisheries.