Background: Although Nivolumab (NIVO) or Irinotecan (IRI) monotherapy is one of the later-line chemotherapies for the advanced gastric cancer (AGC) patients, there are no prospective data of their comparison, and no indicators for selection of them. Methods: Clinical data of AGC patients treated with NIVO or IRI as later-line were collected from 7 institutions between September 2015 and September 2018. Efficacy of the two treatment groups was compared in the entire cohort and in the propensity score matching cohort. Factors which affect the prognosis were explored by multivariate analyses and the scoring system was developed using Cox regression. Results: In the analysis of entire cohort (317 patients), the median OS was 4.8 months (95% CI: 3.6–5.8) in the NIVO group and 5.9 months (95% CI: 4.9–7.2) in the IRI group, and those were 4.8 months (95% CI: 3.6–5.7) and 6.0 months (95% CI: 4.5–7.2) in the analysis of the propensity score matching cohort (248 patients), respectively. There were no differences in OS. We developed a risk score involving 1-Neutrophil to Lymphocyte Ratio (NLR), 2-age, 3-ECOG performance status, 4-peritoneal metastasis, 5-gastrectomy, and 6-liver metastasis to assess potential effect modification. Based on the result, IRI was favored in the higher-risk group (≥4 points, with double weight to 2, 3, and 4) for both OS/PFS. In contrast, NIVO was better OS/PFS in the lower-risk group (<1 point). Conclusions: According to the real-world data from multicenter in Japan, NIVO and IRI generally showed equivalent outcomes. The scoring system we explored may be useful for selecting appropriate later-line chemotherapy, either NIVO or IRI.