ObjectivesOral cancer is the ninth most common cancer worldwide and a leading cause of cancer-related death. Oral squamous cell carcinoma (OSCC) accounts for 90% of all oral cancers. Autophagy is a conserved essential catabolic process related to OSCC. The aim of this study was to elucidate diagnostic and prognostic autophagy-related biomarkers in OSCC. MethodsThe OSCC gene expression data set was obtained from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) between the OSCC samples and adjacent healthy tissues were identified by R software. The Human Autophagy Database was screened, which revealed 222 autophagy-related genes. The autophagy-related DEGs were identified. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were applied. Protein–protein interaction network analysis was performed in the STRING database. cytoHubba in the Cytoscape software was applied to determine the top 10 hub genes. The data set of patients with OSCC from The Cancer Genome Atlas (TCGA) was used to evaluate the prognostic value of the 10 hub genes. The association between prognosis-related hub genes and immune infiltrates was explored. ResultsTwenty-seven autophagy-related DEGs were identified. The top 10 hub genes were CCL2, CDKN2A, CTSB, CTSD, CXCR4, ITGA6, MAP1LC3A, MAPK3, PARP1, and RAB11A. ITGA6 was identified as the most efficient biomarker. Receiver operating characteristic curve analysis indicated that ITGA6 had the highest diagnostic accuracy for OSCC (area under the curve = 0.925). ITGA6 expression was significantly related to immune infiltrates. ConclusionsThe autophagy-related gene ITGA6 might be an efficient diagnostic and prognostic biomarker in OSCC.