The aim of this study was to determine whether skeletal muscle depletion predicts the prognosis of patients with hepatocellular carcinoma (HCC) that is being treated with sorafenib. We evaluated 40 consecutive HCC patients who received sorafenib treatment. The skeletal muscle cross-sectional area was measured by computed tomography at the third lumbar vertebra (L3), from which the L3 skeletal muscle index (L3 SMI) was obtained. The factors contributing to overall survival, sorafenib dose reduction, and discontinuation of sorafenib were analyzed using the Cox proportional hazards model. L3 SMI (p = 0.020) and log (α-fetoprotein (AFP)) (p = 0.010) were identified as independent prognostic factors in HCC patients treated with sorafenib. The initial dose of sorafenib (p = 0.008) was an independent risk factor for sorafenib dose reduction, and log (AFP) (p = 0.008) was the only significant risk factor for the discontinuation of this drug. L3 SMI was not a risk factor for either dose reduction (p = 0.423) or the discontinuation (p = 0.132) of sorafenib. A multiple linear regression analysis determined the following relationship between skeletal muscle mass (assessed as L3 SMI) and the explanatory factors: L3 SMI = -0.1896 × (Age) - 10.3441 × (Child-Pugh score) - 9.3922 × (log (AFP)) + 1.6139 × (log (AFP)) × (Child-Pugh score) + 112.9166. Skeletal muscle depletion is inversely associated with age, Child-Pugh score, and log (AFP). Moreover, it is an independent prognostic factor for HCC patients treated with sorafenib.