Head and neck squamous cell carcinoma (HNSCC) is one of the most common cancers worldwide, accounting for almost 50% of all malignancies in developing nations. Autophagy plays a vital role in cancer initiation, malignant progression, and resistance to treatment. However, autophagy-related gene sets have rarely been analyzed in HNSCC. Hence, it is necessary to assess its clinical and pathological significance in a larger cohort of patients with HNSCC. The purpose of this study was to establish a novel autophagy-related prognostic marker for HNSCC. We screened 232 autophagy-related genes (ARGs) and identified 38 differentially expressed ARGs in The Cancer Genome Atlas (TCGA) cohorts. The prognosis-related ARGs signature, established using the univariate and multivariate Cox proportional regression models, consists of 10 ARGs that could divide patients into high-risk and low-risk groups. Survival analysis indicated that patients in the high-risk group had dramatically shorter overall survival compared with their low-risk counterparts. Cox regression analysis further confirmed the independent prognostic value of the autophagy-related signature, and the area under the receiver operating characteristic curve of the combined prognostic model was 0.722. Finally, the efficacy of autophagy-related signature was also validated by an independent cohort from the Gene Expression Omnibus (GEO) database. Collectively, we successfully constructed a novel autophagy-related signature for the prediction of prognosis in patients with HNSCC.