Background: Esophageal cancer (EsC) often recurs early after surgery. Recent studies shown that commensal bacteria affect the treatment for cancer through the mediation of host immune system and become a biomarker for recurrence of cancer after surgery in colorectal cancer. We, therefore, hypothesize that commensal bacteria are a potential biomarker of postoperative recurrence of EsC. This study leads to clarifying the immunological mechanism of the EsC microenvironment. Method: We collected the stool sample from patients that underwent surgery for esophageal cancer at the Department of General and Gastroenterological Surgery of Showa University, between January 2017 and September 2018. Patients were followed for twelve months to allow time to assess recurrences of EsC. We investigated the composition of the bacterial microbiome in the stool of patients with EsC by using 16s rRNA sequencing. Furthermore, we identified specific microbiotas as a risk factor in the recurrence of EsC by using AI/machine learning. Result: 16s analysis with Qiime2 showed a significant difference between recurrence of EsC group and non-recurrence of EsC group in the bacteria A in the Odoribacteraceae family. AI/machine learning showed a significant difference between the two groups in 8 species including bacteria A in the Odoribacteraceae family and bacteria B in the Actinomyces family. Bacteria A was significant higher in the recurrence group with the two analysis methods. Conclusion: Commensal bacteria become a potential biomarker of postoperative recurrence of EsC. Among commensal bacteria, bacteria A has been suggested to be a biomarker as a risk factor for EsC recurrence.