SUMMARY This study aimed to investigate the relationship among somatic cell scores (SCS) and currently selected traits (milk yield, and fat, protein, lactose, total solids and no-fat-solids contents) in Holstein dairy cows in a semi-arid climate. The data were collected from available records of 272 Holstein cows from January 2007 to December 2016. Farms were located in Agreste region of Pernambuco whose climatic characteristics according to Instituto Nacional de Meteorologia-INMET were: average annual temperature of 25.21 °C; average annual precipitation 562.07 mm, and average annual relative humidity 75.61%.. A total of 4,586 records of somatic cell counts (SCC) in the milk, milk yield and milk components was analyzed. The SCC was transformed logarithmically into SCS SCS = log 2 ( SCC / 100 , 000 ) + 3 ) before statistical investigation. Ten SCS were obtained for determining their effects on milk yield and composition. Data were processed using PROC GLM and PROC CORR procedures in SAS. Results showed positive correlations among SCS and fat, protein, total solids and solids-non-fat contents, while the SCS and lactose content and milk yield were negatively correlated. The highest milk yield (34.43 kg/cow/day) was obtained for the lowest SCS (0; 0 to 24 cells x 1000/mL). The milk yield and lactose decreased while protein and fat contents increased when SCS increased mostly above score five (400 to 799 cells x 1000/mL). It was observed that the increase in SCS influenced negatively milk yield and composition in Holstein cows created in the semi-arid climate in Brazil.