The objective was to compare productive and metabolic responses of lactating dairy cows managed on 2 pasture-based systems using a concentrate supplement (n = 16) with those of a freestall housing system (n = 24). In a 259-d experiment, 3 multiparous Holstein cows were assigned at calving to each of 4 replicates of 2 pasture systems. For system 1, winter pastures were a mixture of rye, ryegrass, and crimson and red clover; summer pastures were pearl millet. Pasture system 2 included a rye-ryegrass mixture during winter and bermudagrass during summer. Pregraze herbage mass averaged 2.3 and 3.6 Mg/ha for winter and summer pastures, respectively; however, during August through September, pearl millet pregraze mass was reduced to about 1 Mg/ha. Daily dry matter intake by cows on pasture averaged 24.7 kg/d in winter and 19.0 kg/d in summer, of which 55% was from pasture; that of cows in confined-housing averaged 23.6 kg/d. Cows in confinement produced 19% more milk (29.8 vs. 25.1 kg/d) than those on pasture systems. Differences in concentration of milk fat, protein, or urea N were not detected among treatment groups. Grazing cows lost more body weight than confined cows (113 vs. 58 kg) and had lower concentrations of plasma glucose in the early weeks postpartum. Despite greater milk yield by cows housed in freestalls, milk income minus feed costs including that of pasture was similar for the 3 management systems. Although these pasture systems might be a viable management system in the southeastern US, extensive loss of body weight immediately postpartum for pasture-based cows are a potential concern.