This study reports the 2-year outcomes and biomarker analysis results of patients with locally advanced gastric and gastroesophageal junction (G/GEJ) adenocarcinoma who received neoadjuvant chemotherapy and immunotherapy in a phase II WuhanUHGI001 trial. Eligible patients with cT3/4aN+M0 locally advanced G/GEJ adenocarcinoma were screened, enrolled, and treated with 3 cycles of neoadjuvant tislelizumab and SOX followed by D2 gastrectomy and another 5 cycles of postoperative adjuvant SOX. The primary endpoint was major pathological response. Of the 49 included patients, 24 (49.0%) achieved major pathological response and 13 (26.5%) achieved pathological complete response. During a median follow-up of 26.8 months, the 2-year progression-free survival (PFS) and overall survival (OS) rates were 69.4% and 81.2%, respectively. Grade 3-4 adverse events occurred in six patients (12.2%) during the neoadjuvant period, eight patients (17.0%) during the postoperative period, and seven patients (15.2%) during the adjuvant period. Biomarker analysis revealed that the pathological complete response showed no association with 2-year PFS and OS. Major pathological response showed a potentially strong association with improved 2-year PFS and OS rates. In addition, preoperative circulating tumor cells combined with pathological responses are helpful in prognosis assessment. In addition, our results showed that T downstaging, lymphocyte-to-monocyte ratio, and CD3+ T cells were independent factors that affect PFS. The signet ring cell component (SRCC), T downstaging, and neutrophil-to-lymphocyte ratio were independent factors affecting OS. Prognostic nomograms of PFS and OS constructed based on the multivariate Cox regression results demonstrated suitable calibration and discrimination ability. Neoadjuvant tislelizumab plus SOX exhibits promising efficacy and acceptable toxicity in patients with locally advanced G/GEJ adenocarcinoma. In addition, our study established a prognostic risk signature and nomograms based on clinicopathological characteristics, which can accurately predict patient outcomes and aid in personalized treatment planning.