Background: Inflammation is a long-established hallmark of liver fibrosis and carcinogenesis. Eosinophils are emerging as crucial components of the inflammatory process influencing cancer development. The role of blood eosinophils in patients with hepatocellular carcinoma receiving systemic treatment is an unexplored field. Objective: The objective of this study was to analyse the prognostic role of the baseline eosinophil count in patients with sorafenib-treated hepatocellular carcinoma. Patients and Methods: A training cohort of 92 patients with advanced- or intermediate-stage sorafenib-treated hepatocellular carcinoma and two validation cohorts of 65 and 180 patients were analysed. Overall survival and progression-free survival in relation to baseline eosinophil counts were estimated by the Kaplan–Meier method. Univariate and multivariate analyses were performed. Results: A negative prognostic impact of low baseline eosinophil counts (< 50*109/L) was demonstrated in all cohorts (training cohort: hazard ratio = 50.1, 95% confidence interval 11.6–216.5, p < 0.0001 for low vs high eosinophil counts; first validation cohort: hazard ratio = 4.55, 95% confidence interval 1.24–16.65, p = 0.022; second validation cohort: hazard ratio = 3.21, 95% confidence interval 1.83–5.64, p < 0.0001). Moreover, low eosinophil counts had a negative prognostic role in patients progressing on or intolerant to sorafenib who received second-line regorafenib, but not capecitabine or best supportive care. Conclusions: Our analysis identified baseline blood eosinophil counts as a new prognostic factor in patients with sorafenib-treated hepatocellular carcinoma. Concerning second-line therapies, eosinophil counts were associated with survival outcomes only in regorafenib-treated patients, suggesting a possible predictive role in this setting.