Productivity in most Brazilian dairy herds is low and depends exclusively on pasture. To study the productive potential of pastures and to devise strategies to further improve pasture and animal productivity in this production system, studies were carried out to obtain basic on-farm information. The constraints which affect productivity and reproductive performance of dairy cows, the effects of restrictions in suckling time of calves, and strategic supplementation during the dry season upon animal production were the evaluated parameters. From March 1992 through February 1997, studies were carried out on four private farms in the northern region of the State of São Paulo. Between March 1992 and February 1994 (Study 1--survey phase), 142 cows (parity = 1-6) grazed pasture which consisted of signal grass (Brachiaria decumbens and Brachiaria brizanta). Once-a-month data were collected on body weight, body condition, and milk production. Reproduction parameters were assessed by milk progesterone profiles. From March 1996 to February 1997 (Study 2--intervention phase), 45 lactating dairy cows from two farms were hand-milked once a day and the calf suckling was restricted to two hours after milking. Data were collected on milk production and cow body weight. In Study 1, cows were grouped by calving date for the analysis of the reproductive and production data. Concentrations of blood metabolites, hemoglobin, and hematocrit were compared among randomly selected cows (n = 69) from all farms. Estimated pasture available per hectare (ha) at any time, crude protein (CP), and dry matter digestibility (DMD) of pasture available for grazing differed (p < 0.05) between seasons [pasture available = 1.2, 1.4, 1.8 and 2.2 t/ha (SE = 0.70); CP = 42, 60, 48 and 57 g/kg (DM) (SE = 10.1); DMD = 399, 468, 401 and 457 g/kg (DM) (SE = 21), respectively, for dry season 1992 (D92), wet season 1992 (W92), dry season 1993 (D93), and wet season 1993 (W93)]. The proportion of animals showing ovarian activity at 90 days postpartum (DPP) was higher for cows which calved in the wet season than cows which calved in the dry season. In Study 2, milk production tended to the higher (7.3 and 6.5 kg/day, respectively, for the intervention and survey studies; p = 0.08). The data suggest that milk production is being limited by pasture availability, the quality of pasture, and the lack of supplementation. We suggest that, although supplementing cows in the dry period may have an economic advantage, better pasture management needs to be introduced. Stocking rate must be adapted to pasture productivity and pasture quality throughout the year.