In order to estimate the genetic effects - direct and maternal - linked to the expression of growth traits in lambs of Junín breed, four mixed models were applied in the study of the birth weight (PN), weaning weight (PD), first shearing weight (PE) and daily body weight gain from birth to weaning (GPND), through univariate and bivariate analysis. The models used in univariate analysis included as fixed effects year of birth and sex, and as covariates the mother's weight at mating (for PN and PD), age of lamb at weaning (for PD and GPND), and age at first shearing (for PE). Model 1 included the fixed effects and the direct additive genetic effect (a); Model 2, similar to 1, added additive maternal genetic effect (m), considering Covam=0; Model 3, similar to 2, assumed Covam=Aσam; and Model 4, equal to 3, added the maternal permanent environmental effect (c). All bivariate models were extensions of the univariate ones. In all the analyses, the restricted maximum likelihood (REML) method, implemented in the ASReml program was applied. Univariate Model 2 and bivariate Model 3 estimated the best genetic parameters, including correlations between direct and maternal additive effects for the traits. The estimated heritability values ranged from low to moderate, indicating the possibility of its application in genetic improvement plans for Junín sheep, which would increase its efficiency when the estimated, moderate to high genetic and phenotypic correlations are also included.