Twenty Angus bulls from three purebred herds in Georgia were acquired to determine the impact of selecting sires based on phenotypic yearling ultrasound intramuscular fat % (UIMF) or UIMF expected progeny difference (EPD) on marbling score of steer progeny managed for commercial slaughter. Each year in each herd, pairs of bulls were selected to create large differences based on their age adjusted phenotypic yearling UIMF measurements. The average UIMF, weighted by number of progeny per sire, was 3.75% (SD = 1.10%) and 1.70% (SD = 0.53%) for high UIMF (HU) and low UIMF (LU) bulls, respectively. All available ultrasound measurements collected in the purebred cooperator herds were combined with other ultrasound records collected by the American Angus Association for the computation of genetic values for ultrasound fat thickness, ribeye area, and intramuscular fat %. Bulls were randomly mated to between 14 and 30 commercial Angus females for 1 to 5 yr. Carcass weight, fat thickness at the 12th rib, ribeye area at the 12th rib, marbling score, yield grade, and quality grade measurements were taken on 188 steer progeny. Carcass data were linearly adjusted to 480 d of age at slaughter. Steer progeny sired by HU bulls had higher age adjusted marbling score and quality grade (P < 0.05), and smaller age adjusted ribeye area (P < 0.05) than progeny sired by LU bulls. The regression of age adjusted carcass marbling score of steer progeny on ultrasound intramuscular fat % EPD of sires produced a highly significant regression coefficient of 90.50. Age adjusted carcass quality grade of steer progeny regressed on ultrasound intramuscular fat % EPD of sires produced a highly significant regression coefficient of 49.20. Yearling Angus bulls selected for high phenotypic UIMF or UIMF EPD can be expected to produce steers with significantly higher amounts of marbling and quality grade. It also appears that marbling can be increased without corresponding increases in external fat thickness and yield grade.