The achievement of pathological complete response (pCR) with neoadjuvant therapy can significantly improve prognosis in patients with gastric cancer (GC). GC tissues demonstrating pCR after immunotherapy exhibited increased stiffness and proliferation of fibroblasts within the stroma. Specific subpopulation cancer-associated fibroblasts (CAFs) may serve as potential markers for predicting the efficacy of immunotherapy. We screened CAFs-related genes as candidate predictors for immunotherapy using the TCGA-STAD, PRJEB25780, GSE27342, and GSE54129 databases. Tissue specimen from GC patients enrolled in the clinical trial (NCT04208347) was used to evaluate its clinical significance. Single-cell RNA sequencing (scRNA-seq) data were obtained from GSE163558, GSE183904, and GSE184198 datasets and analyzed through Seurat v3 R software and iTALK. GC patient-derived organoids (GC-PDOs) modeling verified the effect of CAFs subpopulations on immunotherapeutic response . Podoplanin (PDPN) has been identified as a candidate marker related to CAFs for predicting the efficacy of immunotherapy. Western blot analysis indicated that lower PDPN expression was observed in GC samples with pCR. Functional and pathway enrichment analysis indicated PDPN was associated with numerous malignancy-related pathways in gastric cancer. Using the iTALK algorithm, scRNA-seq datasets further verified the interaction between a subpopulation of PDPN+ CAFs and immune cells. The results of multiple immunohistochemistry/immunofluorescence suggested a negative correlation between PDPN+ CAFs and pCR to anti-PD-1 treatment ( < 0.01). Notably, using the GC-PDO model, we determined that PDPN + CAFs hinder the activation, thereby reducing immune response in GC patients. PDPN+ CAFs subpopulation has a potential correlation with the efficacy of immunotherapy for GC patients.