Various serum biomarkers have been developed for predicting head and neck squamous cell carcinoma (HNSCC) prognosis. However, none of them have been proven to be clinically significant. A recent study reported that the ratio of aspartate aminotransaminase (AST) to alanine aminotransaminase (ALT) had a prognostic effect on non-metastatic cancers. This study aimed to examine the effect of the AST/ALT ratio on the survival of patients with HNSCC. Clinical data of 356 patients with locoregionally advanced HNSCC were collected. The effect of the AST/ALT ratio on overall survival was analyzed using a Cox proportional hazard model. Moreover, recursive partitioning analysis (RPA) was used to divide the patients into groups on the basis of the clinical stage and AST/ALT ratio. The prognostic ability of this grouping was validated using an independent data set (N = 167). The AST/ALT ratio ranged from 0.42 to 4.30 (median, 1.42) and was a prognostic factor for overall survival that was independent of age, primary sites, and tumor stage (hazard ratio: 1.36, confidence interval: 1.08-1.68, P = 0.010). RPA divided patients with stage IVA into the following two subgroups: high AST/ALT (≥2.3) and low AST/ALT (<2.3) subgroups. The 5-year survival rate for patients with stage III, stage IVA with a low AST/ALT ratio, stage IVA with a high AST/ALT ratio, and stage IVB were 64.8%, 49.2%, 28.6%, and 33.3%, respectively (p < 0.001). Compared with the low AST/ALT group, the adjusted hazard ratio for death was 2.17 for high AST/ALT group (confidence interval: 1.02-.22 P = 0.045). The AST/ALT ratio was demonstrated to be a prognostic factor of HNSCC. The ratio subdivided patients with stage IVA into low- and high-risk groups. Moreover, intensified treatment for the high-risk group may be considered.