Weaning weight in a single flock of 600 ewes and 40 rams was simulated through 30 consecutive lambings. The objective was to compare the simple breeding value model with the maternal effects model for lamb evaluation under the Segureña selection scheme. Three selection strategies were tested: selection on breeding values estimated by a simple breeding value model that ignores maternal effects (method 1), selection on direct additive values only (method 2) or on the sum of direct and maternal additive values (method 3), both latter methods utilizing a maternal effects model. Average values obtained in the last lamb crop were about 18 kg for the phenotypic mean, about 5 kg for direct additive values and about -1 kg for maternal additive values, for all three methods. Average inbreeding coefficient of the last crop was more than 0·13 in method 1 but was less than 0·12 in the others. All differences were not statistically significant (P > 0·05). Consequently, the simple breeding value model can be used for the genetic evaluation of weaning weight of candidates for selection in the Segureña scheme. The effect of variation in the magnitude of parameters was evaluated through four sets of parameters. Results showed that with a higher additive maternal component, method (3) would become increasingly necessary. The need for method (3) is accentuated with more negative additive covariance between direct and maternal effects. Thus, for higher maternal effects or more negative additive correlation, the use of the complete model (with maternal effects) becomes unavoidable. Varying population size, however, affects only the inbreeding accumulated, as long as the same methods of selection are used.