The performance of steers raised on elephant and mombasa grass pastures was evaluated from October to December 2006. A rotational grazing system with a regulated forage offer was adopted. The experimental area consisted of 2 block containing 18 paddocks of 0.25 ha divided into 9 paddocks with elephant grass and 9 with mombasa grass. Forage species, area block, the interaction among these effects, paddocks within the interaction, and the experimental error were the effects of the treatments. This model was used to analyze pasture attributes. For analyses of intake, digestibility and gain, 24 steers were randomly assigned to the respective interactions among forage vs. block vs. grazing system, totaling 3 animals per triple interaction. The grazing systems consisted of ad libitum (AL) and restricted (PR) grazing. Elephant-grass showed the highest (P < 0.05) total dry matter biomass (TB) and green leaf dry matter biomass (GLDM). There were no differences (P > 0.05) among forage plants for production of biomass of green leaf dry matter; however, leaf proportion was the greatest (P < 0.05) in mombasa grass. Extrusas showed similar (P > 0.05) bromatological composition. Only crude protein intake, which was the highest for mombasa pasture, differed (P < 0.05) among forage plants. Ad libitum feeding enabled a higher (P < 0.05) intake of all nutrients from both pastures. Digestibility of nutrients was higher for elephant-grass but it did not differ (P > 0.05) among feeding levels:restricted grazing allowed for greater plant crude protein concentration while ad libitum grazing resulted in greater gross energy consumption. In conclusion, despite qualitative and quantitative differences among forage plants, differences of weight gain per animal and per area were not found. For animal with ad libitum feeding, average body mass gain was 850 g/day and gain per area was 246 kg/ha during the experimental period. Funded by CAPES, CNPq, FAPEMIG, and FAPERJ.