Background: Nivolumab has changed the treatment of advanced gastric cancer (AGC). Nivolumab shows better outcome compared to best supportive care in AGC patients who received at least two prior regimen. Although there is not reliable date of poor performance status(PS) AGC patients who received nivolumab. We investigated efficacy and safety of nivolumab for AGC patients with poor PS. Methods: We retrospectively collected clinicopathologic data from patients with AGC who received nivolumab monotherapy in Himeji Red Cross Hospital from October 2017 to June 2019. Results: 49 AGC patients who received nivolumab were analyzed. 27 patients were PS 0 or 1(Good Group), and 22 patients were PS 2 or 3(Poor Group). Median progression free survival and overall survival was 61 days and 180 days in Good Group and 36 days and 85 days in Poor Group. Overall survival (OS) was significantly shorter in Poor group(180 days vs 85 days, p = 0.0255). Disease control rate was 23% in Good group and 9% in Poor group. 33% patients were experienced immune related adverse event (iRAE) in Good Group, and 18% in Poor Group. We investigated prognostic factor of OS in Poor Group such as Royal Marsden Hospital Score(RMH score), modified Glasgow prognostic score(mGPS), and Japan Clinical Oncology Group (JCOG) prognostic index. RMH score and JCOG prognostic index good or moderate group was significantly longer overall survival than poor group (93 days vs 35 days (p = 0.0214)). JCOG prognostic index was most correlated with OS among these tools. Conclusions: This study suggested that nivolumab has a modest effect and is feasible as third line or later line for AGC patients. JCOG prognostic index was suggested to be effective in predicting prognosis in AGC patients who received nivolumab.