Quantifying DMI is necessary for estimation of nutrient consumption by ruminants, but it is inherently difficult on grazed pastures and even more so when supplements are fed. Our objectives were to compare three methods of estimating forage DMI (inference from animal performance, evaluation from fecal output using a pulse-dose marker, and estimation from herbage disappearance methods) and to identify the most useful approach or combination of approaches for estimating pasture intake by lactating dairy cows. During three continuous 28-d periods in the winter season, Holstein cows (Bos taurus; n = 32) grazed a cool-season grass or a cool-season grass-clover mixture at two stocking rates (SR; 5 vs. 2.5 cows/ha) and were fed two rates of concentrate supplementation (CS; 1 kg of concentrate [as-fed] per 2.5 or 3.5 kg of milk produced). Animal response data used in computations for the animal performance method were obtained from the latter 14 d of each period. For the pulse-dose marker method, chromium-mordanted fiber was used. Pasture sampling to determine herbage disappearance was done weekly throughout the study. Forage DMI estimated by the animal performance method was different among periods (P < 0.001; 6.5, 6.4, and 9.6 kg/d for Periods 1, 2, and 3, respectively), between SR (P < 0.001; 8.7 [low SR] vs. 6.3 kg/d [high SR]) and between CS (P < 0.01; 8.4 [low CS] vs. 6.6 kg/d [high CS]). The period and SR effect seemed to be related to forage mass. The pulse-dose marker method generally provided greater estimates of forage DMI (as much as 11.0 kg/d more than the animal performance method) and was not correlated with the other methods. Estimates of forage DMI by the herbage disappearance method were correlated with the animal performance method. The difference between estimates from these two methods, ranging from -4.7 to 5.4 kg/d, were much lower than their difference from pulse-dose marker estimates. The results of this study suggest that, when appropriate for the research objectives, the animal performance or herbage disappearance methods may be useful and less costly alternatives to using the pulse-dose method.