The clinical impact of neutrophil-to-lymphocyte ratio (NLR) in predicting outcomes in hepatocellular carcinoma (HCC) patients treated with transarterial chemoembolization (TACE) remain unclear, and additional large-scale studies are required. This retrospective study evaluated outcomes in treatment-naïve patients who received TACE as first-line treatment for intermediate-stage HCC between 2008 and 2017. Patients who underwent TACE before and after 2013 were assigned to the development (n = 495) and validation (n = 436) cohorts, respectively. Multivariable Cox analysis identified six factors predictive of outcome, including NLR, which were used to create models predictive of overall survival (OS) in the development cohort. Risk scores of 0–3, 4–7, and 8–12 were defined as low, intermediate, and high risk, respectively. Median OS times in the low-, medium-, and high-risk groups in the validation cohort were 48.1, 24.3, and 9.7 months, respectively (p < 0.001). Application to the validation cohort of time-dependent ROC curves for models predictive of OS showed AUC values of 0.72 and 0.70 at 3 and 5 years, respectively. Multivariable logistic regression analysis found that NLR ≥ 3 was a significant predictor (odds ratio, 3.4; p < 0.001) of disease progression 6 months after TACE. Higher baseline NLR was predictive of poor prognosis in patients who underwent TACE for intermediate-stage HCC.