There is currently a lack of validated predictors for adjuvant chemotherapy efficacy in patients with gastric cancer (GC). Perineural invasion (PNI) is the process of neoplastic invasion of the nerves, accompanied by tumor microenvironment (TME) changes. TME can determine treatment outcome while the impact of PNI on chemotherapy efficacy remains unknown in GC. We investigated the association between PNI and the efficacy of postoperative adjuvant chemotherapy in patients with resected GC. Patients who underwent radical resection of stage IB-III GC with or without fluoropyrimidine (FU)-based adjuvant chemotherapy were retrospectively selected from two separate patient cohorts. PNI was confirmed with S100 immunohistochemistry (IHC). Tumor hypoxia and activity of selected pathways were quantified by mRNA-based signature scoring based on publicly available data. A hypoxia biomarker, ERO1A, and a FU resistance biomarker, thymidylate synthase (TS), were assessed by IHC, respectively. Two cohorts included 223 and 599 patients, respectively. Adjuvant chemotherapy significantly improved overall survival (OS) and disease-free survival (DFS) in PNI-positive but not in PNI-negative patients, which was not impacted by stages. Multivariate models showed that adjuvant chemotherapy was an independent predictor for OS and DFS in PNI-positive patients in both cohorts. For TME, PNI-negative tumors were more hypoxic than were PNI-positive tumors, and displayed relative up-regulation of signaling along the pathways that are important in FU metabolism or resistance. Expressions of ERO1A and TS significantly decreased in PNI-positive compared to PNI-negative tumors. PNI might help predict adjuvant chemotherapy benefit in patients with resected GC. Validation in prospective studies is required. Novel treatment strategies need to be developed in PNI-negative patients.