Dry‐matter intake (DMI) is difficult to measure. The potential of field‐based near‐infrared (NIR) proximal sensing as a means to quantify DMI is a logical extension of the findings of several researchers who demonstrated that benchtop NIR spectroscopy could be used to predict animal intake. Our objectives were to predict DMI from the relative feed value and the Relative Forage Quality equations and to compare these estimates to that measured in lambs. Proximally, sensed spectra of live forage were used to develop predictions of neutral detergent fibre, acid detergent fibre, in vitro true digestibility (IVTD) and crude protein which were used as inputs to the DMI equations. These prediction equations accounted for 68%, 87%, 84% and 71% of the variation in their measured counterparts, and when used in the DMI equations, accounted for about 80% and 79%, respectively, of the variation of that calculated using laboratory‐measured values. However, neither the laboratory‐ nor proximally sensed estimates of DMI were correlated with that measured DMI in the lambs. Proximally sensed IVTD was incorporated into a sheep‐based intake model where from 7% to 36% of the variability in measured DMI, depending on breed, was explained. This improvement in estimated DMI suggests: 1) that DMI equations based only on forage nutritive values are unlikely to provide acceptable estimates, and 2) that proximally sensed forage nutritive values incorporated into a more comprehensive model, such as that described by Freer, et al. (1997, Agric Syst, 54, 77) may provide a timely grazing decision support system.