The increase of sheep meat competitiveness in international markets can be attributed to the rise of the quantity and the improvement of the quality of the edible portion of sheep carcasses. Usually, carcass yield is established after the slaughter of the animals. Yet, when carcass yield is determined in vivo, it can be both a costly and subjective method. This study proposes models for predicting the physical characteristics of lamb carcass using bioimpedance analysis (BIA) in live animals. Thirty-one Texel × Ile de France crossbreed ram lambs were slaughtered at 20, 26, 32 or 38 kg of BW. Before the slaughter, values of resistance (Rs) and reactance (Xc) were collected using a single-frequency BIA equipment (Model RJL Quantum II Bioelectrical Body Composition Analyzer). Then, BIA main variables such as body bioelectrical volume (V), phase angle (PA), resistive density (RsD) and reactive density (XcD) were calculated. After slaughter, cold carcass weight (CCW), cold carcass yield (CCY), subcutaneous fat thickness (SFT), soft tissue weight (STW) and soft tissue yield (STY) were also measured. Multiple regression analyses were carried out using the physical characteristics as dependent variables and the bioimpedance values as independent variables. Predictive performance of the models was assessed using leave-one-out cross-validation. The prediction model of CCW was obtained using the V, PA and RsD (R2 = 0.97), STW through the V, RsD and XcD (R2 = 0.97), CCY by Rs, Z and XcD (R2 = 0.69), STY by V and XcD (R2 = 0.67), and SFT only for XcD (R2 = 0.84). The results indicated that BIA has the potential to predict carcass characteristics of lambs at different body masses.