Genetic and phenotypic correlations between production traits, selected linear type traits, and somatic cell score were estimated. The results could be useful for breeding programs involving Czech Holstein dairy cows or other populations. A series of bivariate analyses was applied whereby (co)variance components were estimated using average information (AI-REML) implemented via the DMU statistical package. Chosen phenotypic data included average somatic cell score per a 305-day standard first lactation as well as the production traits milk yield, fat yield, protein yield, fat percentage, and protein percentage per the standard first lactation. Fifteen classified linear type traits were added, as they were measured at first lactation in the Czech Holstein population. All phenotypic data were collected within the progeny testing program of the CzechMoravian Breeders Corporation from 2005 to 2009. The number of animals for each linear type trait was 59 454, except for locomotion, for which 53 424 animals were recorded. The numbers of animals with records of milk production data were 43 992 for milk yield, fat percentage, protein percentage, and fat-to-protein percentage ratio and 43 978 for fat yield and protein yield. In total, 27 098 somatic cell score records were available. The strongest positive genetic correlation between production traits and linear type traits was estimated between udder width and fat yield (0.51 ± 0.04), while the strongest negative correlation estimated was between body condition score and fat yield (−0.45 ± 0.03). Other estimated correlations were between those two extremes but generally they were close to zero or positive. The strongest negative phenotypic correlations were estimated between udder depth and milk yield and protein yield (both −0.17), while the strongest positive phenotypic correlations were estimated between milk yield, protein yield, and udder width (both 0.32).