Several external markers can be used for estimating total faecal output in view of assessing ruminant intake at pasture. Among them, ytterbium (Yb) has been used for many years in various conditions. Polyethylene glycol (PEG) is a promising external marker because it can be rapidly determined using near-infrared spectroscopy (NIRS). The study consisted of 24 adult lactating dairy ewes over three periods (P1, P2 and P3), fed with three different diets: P1, total mixed ration (TMR); P2, Italian ryegrass (IRG); and P3, pasture. After an adaptation period, the ewes were administered a daily dose of ytterbium oxide (0.35 g/day) and PEG (20 g/day) for 2 weeks. During the last week, the daily organic matter intake (OMIOBS) was measured. Faecal samples were collected at milking time (0800 and 1600 h) to determine marker content, using only samples collected in the morning (PEGm) or by averaging samples (Yb, PEGma). Faecal marker content made it possible to assess total faecal output, either using the two recovery rates for PEG (0.98 or 0.87) or not. The OMIOBS was assessed on the basis of total faeces estimated with Yb (OMIYb) or PEG (OMIPEG), and the digestibility was calculated on the basis of feed analysis. With total TMR (P1), the OMIPEG, corrected with recovery rate (OMIPEGm98) or not corrected (OMIPEGm) was 2.40 kg/day and 2.50 kg/day, respectively, and was not different (P>0.05) from OMIOBS (2.51 kg/day), whereas OMIYb was lower (2.14 kg/day) (P<0.001). With IRG (P2), OMIPEGm98 (1.67 kg/day), OMIPEGm87 (1.51 kg/day) and OMIYb (1.59 kg/day) were not different (P>0.05) from OMIOBS (1.57 kg/day). With pasture (P3), the OMIPEGm (1.54 kg/day) and OMIPEGm98 (1.48 kg/day) were not different (P>0.05) from the OMI assessed from the biomass measurement (1.52 kg/day). The OMIYb (1.36 kg/day) was lower (P<0.05) but not different from OMIPEGm98 and OMIPEGm87. Spearman’s rank correlation between OMIOBS and other OMIs (predicted with Yb or PEG P1 and P2) showed that it is possible to rank animals using PEG when there is a sufficiently wide range of OMIOBS (1.65 to 2.8 kg/day in P1) but not within a narrower range (1.47 to 1.72 kg/day in P2). In conclusion, the present study confirms that PEG is a valuable external faecal marker, easy to prepare (solution), administer and determine (NIRS). It can be used to assess intake with numerous animals at pasture, but only for groups, and not to quantitatively estimate individual OMI.