In close reproductive lines of rabbits reared in discrete generations, there is a strong association between the year-season (AE) and the levels of consanguinity (F). Previous studies have shown that when AE and F were considered as fixed effects in a model, the heritability and the genetic tendency were overestimated. The objective of this work was to investigate the consequences of considering AE as either a fixed (M1) or random (M2) effect on the selection of animals. A total of 15,671 records corresponding to the number of rabbits weaned in line A, from the Department of Animal Science of the Polytechnic University of Valencia, Spain, were used. The comparison of models was based on the rankings of the estimated additive genetic values in each generation and the expected response to the selection. Six scenarios were proposed, varying in selection intensities. In each one, the Spearman correlation was estimated and the percentage of discordance (D) between the animals chosen for M2 but not for M1 was calculated. The response to selection for the animals chosen by either M1 or M2, was estimated as the slope of the regression line between the means of the genetic values predicted by M2 through generations. As the selection pressure decreased, the Spearman correlations increased, decreasing the D. However, at the generation level, both estimators did not always maintain that relationship. The estimated responses to selection were similar when the selection was based on estimated breeding values obtained by using either M1 or M2. Therefore, no differences are expected between both models with respect to selection.