Reliable and flexible methods to assess the nutrition of grazing animals are needed for better management of livestock and natural grasslands. We aimed at comparing two methods of predicting basic diet characteristics of Creole heifers: organic matter digestibility (OM), OM intake (OMI) and digestible OMI (DOMI), by using two methods based on the faecal CP content, the analysis of fecal samples, by a near-infrared spectrometer (NIRS). The comparison of methods was made with data coming from 2 independent and simultaneous trials in 2 feeding systems with Digitaria sp. (offered in stalls, or directly grazed) with two groups of 4 heifers per feeding system, being supplemented or not, during 4 experimental periods (n=64). In a first step, the estimates obtained from the two methods were compared to the in vivo measurements in stall-fed conditions, of OMD, OMI and DOMI. The best prediction of OMD was obtained from the NIRS method (RSD=0.03, R2=0.61, n=32). The best predictions of OMI and DOMI were obtained considering the OMD directly predicted by the NIRS method and the fecal output (FO). Consideration of the concentrate had no significant effect on any of the regression parameters. In a second step, the estimated values of DOMI by the 2 methods were compared to live weight gain (LWG) measured both in stalls and at pasture. In stalls, the LWG predicted with DOMI estimated via the fecal CP content method has provided the lower RSD of prediction to assess the LWG in stalls (RSD=0.74 g/kg LW, R2=0.68, n=32). At pasture, the best prediction of LWG has been predicted by the DOMI directly estimated from the NIRS method (RSD=0.90 g/kg LW, R2=0.53, n=32). Thus, fecal NIRS provided the most reliable estimates of OMD. This OMD combined with FO, provided the most accurate values of OMI and DOMI, compared to in vivo values measured in stalls. However, due to the difficulty to recover the feces at pasture and measure FO, direct predictions by using NIRS give accurate estimates for OMD and satisfactory for OMI and DOMI, consistent with the LWG achieved in stalls as at pasture. This method may therefore contribute to better management of grazing animals.