This study aims to establish a patient-derived tumor xenograft (PDX) model for gastric cancer and apply it in pharmacodynamic studies for chemotherapeutic agents. So as to provide guidance for personalized treatment in gastric cancer patients. Fresh gastric cancer tissues were collected from surgical patients and transplanted into NOG mice to establish PDX model. Human source markers were identified by immunofluorescence and immunohistochemistry methods. Exon sequencing was used to verify the retention of humanized features in the PDX model. The efficacy of six chemotherapeutic drugs (oxaliplatin, cisplatin, paclitaxel, fluorouracil, Tegafur, capecitabine) and optimal dose of capecitabine were evaluated by the PDX model. The study has successfully established the PDX model of gastric cancer with transplantation success rates of P0 to P1 (45%), P1 to P2 (88.89%), and P2 to P3 (87.5%). Importantly, original tumor features were effectively preserved till P3. Drug screening results revealed notable antitumor effect of capecitabine in the PDX model of gastric cancer. The study has successfully established a gastric cancer PDX model, highlighting its potential for chemotherapy drug screening and personalized treatment. The transplantation success rates and preservation of original tumor features underscore the model's reliability and relevance for future studies. Findings from drug screening, especially capecitabine's effectiveness, suggest a promising avenue for precision treatment strategies in gastric cancer patients.