The objective of this study was to quantify the effect of somatic cell score (SCS) on milk yield of dairy cattle. Collection of Dairy Herd Information cow records included: herd, days in milk (DIM), parity, birth date, calving date for each parity, and SCS. Records were collected from 6 states in the Southeast (Tennessee, Kentucky, Virginia, Georgia, Florida, and Mississippi). To be included in the data set, cows must have been within one standard deviation of the mean age of the parity group and have DIM between 240 and 305 d. Herds containing less than 5 cows that met the criteria were not included in the analysis. As a result, 10,333, 5,917, 3,213, and 3,266 cows were included from 468, 357, 238, and 244 herds in the analysis of parities 1, 2, 3, and 4 respectively. The GLM procedure of SAS version 9.3 (SAS Institute Inc. Cary, NC) was used to analyze the following model: Y_ijklmn = S_{i ­}+ H_j + M_k + D_l + A_m + I_n + E_ijklmn, where Y_ijklmn = lactation milk yield, S_i = effect of the ith state, H_j = effect of the jth herd, M_k = effect of kth calving season, D_l = effect of the lth days the cow was in lactation, A_m = effect of the mth age at calving, I_n = effect of nth SCS, and E_ijklmn = residual error. Analysis of parities 1 to 4 occurred separately. Herd was a random effect in the model; all other variables were considered fixed. For every 1-unit increase in SCS, a milk yield loss of −113.58, −209.89, −214.42, and −252.09 kg/lactation occurred for parities 1, 2, 3, and 4 respectively. Overall, first-parity cows have half of the milk production loss from a unit increase in SCS than older parity cows. The Southeast Quality Milk Initiative project is supported by Agriculture and Food Research Initiative Competitive Grant no. 2013-68004-20424 from the USDA National Institute of Food and Agriculture.