Background: Autophagy, is a metabolic pathway that occurs in eukaryotic cells and regulated by autophagy-related genes (ARGs).The occurrence and development of many diseases are caused by abnormal autophagy. The purpose of this article is to explore the relationship between autophagy and prognosis of oral cancer, hoping to provide a new way for early diagnosis and guide doctors to make subsequent treatment decisions.Methods: Download the RNA seq and clinical features of 305 oral cancer and 30 non-tumor patients from The Cancer Genome Atlas (TCGA) dataset. Filtered out differential expression autophagy-related genes (ARGs),and gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyzed these ARGs. Cox regression analysis filtered out the prognostic ARGs and constructed a risk score models for overall survival (OS) .Divided patients into high-risk and low-risk groups based on median risk score. Kaplan-Meier analyzed the overall survival (OS). Next, receiver operating characteristic (ROC) curve verified the predictive accuracy of the model. Furthermore, we performed stratification analyses to explore the relationship between the prognostic signature and clinicopathological variables. Lastly, we used another date set to verify the model. All data was processed by R (version 3.6.0) and perl (version5.18.4).Results: The K-M plot showed the overall survival rate of the high-risk group was lower than the low-risk group’s (P=2.216e−10). And Cox regression analysis suggested that the autophagy prognostic index was an independent prognostic factor. Further more, the ARGs prognostic model was confirmed in dataset of GSE65858.Conclusion: This study constructed an autophagy-related signature of oral cancer, which can foresee the prognosis of patients. It will open up new prospects for fight against oral cancer.