Peritoneal metastasis following gastric cancer surgery is often associated with a poor prognosis. This study aimed to investigate the mechanisms underlying peritoneal metastasis and to develop a predictive model for the risk of postoperative peritoneal metastases in gastric cancer. We performed a comprehensive analysis of the protein mass spectra and tumor microenvironment in paraffin-embedded primary tumor sections from gastric cancer patients, both with and without postoperative peritoneal metastases. Using proteomic profiling, we identified 9595 proteins and stratified patients into three distinct proteomic subgroups (Pro1, Pro2, Pro3) based on differential protein expression. Simultaneously, immune cell profiling allowed us to classify patients into four immune subgroups (IG-I, IG-II, IG-III, IG-IV). The relationships between these proteomic, immune, and metastasis classifications were further explored to uncover potential associations and mechanisms driving metastasis. Building on these insights, we developed an integrative model combining proteomics, immunological, and radiomics data for predicting postoperative peritoneal metastases. This model demonstrated high predictive efficacy, offering a robust tool for identifying high-risk patients. Our findings provide a deeper understanding of the biological processes underlying peritoneal metastasis in gastric cancer, highlighting the interplay between proteomic and immune factors. By establishing novel patient subgroups and an effective prediction model, this study lays the groundwork for early diagnosis and tailored therapeutic strategies to improve outcomes for gastric cancer patients.