Background <p>Frailty is highly prevalent among older gastric cancer patients and strongly associated with adverse postoperative outcomes. While frailty interventions have been implemented, few have focused on family cohesion as a pivotal aspect. This study aims to evaluate the effectiveness of a family cohesion-centered frailty intervention grounded in the Satir Family Therapy Model among for high-risk frailty older patients with gastric cancer.</p> Methods <p>A non-randomized, time-based controlled trial was conducted in a tertiary hospital in China, with 79 participants and their family caregivers allocated according to sequential admission periods into an experimental group (<i>n</i> = 40) or a control group (<i>n</i> = 39). The intervention focused on enhancing family communication and cohesion. Frailty, family cohesion, self-efficacy, anxiety, depression, and quality of life were assessed on the day of admission (T0), at the end of the intervention, specifically at 4&#xa0;weeks after surgery (T1), and 1&#xa0;month after the completion of the intervention (T2). Analyses used generalized estimating equations with multiple imputation under an intention-to-treat framework, adjusting for demographic and clinical covariates. Moderation analyses tested three-way interactions.</p> Results <p>Significant group by time interactions were observed for frailty, family cohesion, self-efficacy, anxiety, depression, and quality of life (all <i>P</i> &lt; 0.05). Compared with usual care, the experimental group showed greater reductions in frailty, anxiety, and depression and was associated with improvement in family cohesion, self-efficacy, and quality of life, with effects strengthening at T2. Sensitivity analyses yielded consistent results. Limited moderating effects were identified for age, education, and TNM stage.</p> Conclusions <p>The family-cohesion-centered intervention was associated with multidimensional and sustained improvements among older gastric cancer patients at high risk of frailty. By integrating the Satir Family Therapy Model into perioperative care, this program may represent a feasible family-based approach that could contribute to optimizing frailty management and postoperative recovery. However, the non-randomized, time-based controlled design may introduce potential biases, including from secular trends or contamination; such biases could affect the estimates in either direction. Further randomized controlled trials are needed to verify this.</p> Trial registrations <p>ChiCTR2400082744.</p>