Many current wild turkey (Meleagris gallopavo) harvest models assume density-independent population dynamics. We developed an alternative model incorporating both nonlinear density-dependence and stochastic density-independent effects on wild turkey populations. We examined model sensitivity to parameter changes in 5% increments and determined mean spring and fall harvests and their variability in the short term (3 yr) and long term (10 yr) from proportional harvesting under these conditions. In the long term, population growth rates were most sensitive to poult:female ratios and the form of density dependence. The nonlinear density-dependent effect produced a population that maximized yield at 40% carrying capacity. The model indicated that a spring or fall proportional harvest could be maximized for fall harvest rates between 0% and 13% of the population, assuming a 15% spring male harvest and 5% spring illegal female kill. Combined spring and fall harvests could be maximized at a 9% fall harvest, under the same assumptions. Variability in population growth and harvest rates increased uncertainty in spring and fall harvests and the probability of overharvesting annual yield, with growth rate variation having the strongest effect. Model simulations suggested fall harvest rates should be conservative (≤9%) for most management strategies.