Genetic parameters were estimated by multitrait REML programme using an animal model for udder health traits [clinical mastitis (MAST) and somatic cell counts (SCC)] and milk yield traits [305day milk yield (MY), 305day fat yield (FY) and 305day protein yield (PY)] in the first three lactations of 4015 Friesian cows records calved first from 2000 to 2005 at Sakha farm. The mixed model used in the analysis included the fixed effects of month and year of calving, and parity and the random effects of additive genetics, maternal permanent environment and residual. For each lactation, heritability estimates for MAST (0.13 to 0.20) were slightly higher than those for SCC (0.08 to 0.15) and were from 0.21 to 0.35 for milk yield traits. Heritability estimates for all lactations were 0.21, 0.11, 0.33, 0.29 and 0.29 for MAST ,SCC ,MY ,FY and PY ,respectively. The genetic correlations between MAST and SCC being 0.72 for 1^st , 0.79 for 2^nd  and 0.83 for 3^rd lactations indicated that both traits were genetically favorably associated and selection for low SCC will reduce the incidence of mastitis. The genetic correlations between udder health traits and milk yield traits were mildly positive (0.27 to 0.55) confirming a genetic antagonism between production and udder health traits in this herd. Genetic correlations for each trait across lactations were positive and high for all traits studied, MAST (0.70 to 0.93), SCC (0.84 to 0.91) and milk traits (0.88 to 0.99). The strongest correlations were between second and third lactation for all investigated traits. The positively high genetic correlations between both MAST and SCC in subsequent lactations suggested that susceptibility to mastitis remains unchangeable by advanced of age. For all lactations, genetic correlations were slightly lower than subsequent lactations. Estimates of permanent environmental variances as proportions of phenotypic variances were, 0.09, 0.39, 0.11, 0.22 and 0.23 for MAST, SCC, MY, FY and PY, respectively. Selection of the cow on the basis of their expected breeding values should cause reduction in the incidence MAST and SCC, and increase MY, FY and PY. The accuracy of the index that included SCC and MAST with MY was about 7 times higher than that of other indices. It is emergent to include udder health traits in the breeding programme by using SCC as a criterion for selection for mastitis resistance.