A deterministic computer simulation model was used to estimate the effects of different culling strategies for nonconception and age on the biological efficiency of an annual lambing system. Interaction among culling strategy and genotype for system efficiency was estimated by also varying ewe genetic potentials for fertility, precocity of fertility, and litter size. Biological efficiency was defined as grams of empty body weight equivalent value of market lamb, cull ewe, and wool output per kilogram input of TDN. Pure breeding with random selection of intraflock replacement ewe lambs was assumed. Sheep were fed to maintain normal weight. Maximum allowable ewe age (i.e., time at culling for age) was varied from 2 to 10 yr. The four strategies of culling for nonconception that were simulated ranged from no removal for nonconception to removal of all nonpregnant ewes 15 wk after exposure. During calculation of lamb empty body weight equivalent, relative values per kilogram of output for market lambs, cull ewes, and clean wool were assumed to be 1 to .33 to 2.04 for most simulation runs. In addition, relative value of cull ewes was varied to .01, .16, .50, or .99 for some simulations. The effect of strategy for removal of nonconceiving ewes on system efficiency depended on maximum ewe age and genotype. In general, if relative value per kilogram of cull ewes to market lambs was < or = .50, ewes should be allowed to produce through 4 or 5 yr of age and then be salvaged. Unless salvage value of culled ewes approaches that of market lambs, ewe lambs should not be culled for failure to conceive.