This study was conducted in a forest dominated by 'ñandubay' Prosopis affinis and 'espinillo' Acacia caven located in Feliciano, Entre Rios, Argentina, to implement rotational grazing with high bovine density, evaluating the time of grazing by means of forage production and grass response. Such work was a technical advising whose target was to carry an experimental test of 473 ha of native forest, divided in 6 plots with rotational grazing in an agricultural establishment with a pastoral base of 9000 ha. The implementation of rotational grazing began in March 2002. The forage availability was calculated by the number of allowances and time of grazing for each plot according to the season, from August 2002 to January 2003. To calculate the dry forage biomass, 10 samples of 0.5 m² were placed when entering the cows to grazing. The herbaceous level was described and the vegetable cover was studied by productive types (Rosengurtt, 1979). An average of 51 allowances/ha/day was estimated with a maximum of 100 and a minimum of 14 allowances/ha/day. The results showed that for instantaneous stocking rate and the analysed study period, the best average time of grazing was 15±6 days (maximum 27 and minimum 8) during winter and 13±5 days (maximum 23 and minimum 4) during spring. The real grazing times were different from those estimated due to 2002 rainfall. However, the method proved to be valid to implement rotational grazing due to its capacity to predict the number of grazing days, what allows an acceptable recovery of grass and a proper utilization of forage biomass.