BACKGROUND Anxiety and depression are common among people living with and beyond cancer and frequently recur following the completion of psychological therapy. Digital interventions may offer a scalable and cost-effective means of supporting relapse prevention, but evidence in cancer populations remains limited. OBJECTIVE This study evaluated the feasibility, acceptability, and potential effectiveness of a personalized relapse prevention planning intervention supported by personalized smart-messaging following discharge from psychological therapy for cancer-related anxiety and depression. METHODS A naturalistic, non-randomized quasi-experimental cohort study was conducted within a National Health Service cancer psychology service in England. Patients completing psychological therapy could opt to receive a relapse prevention planning session followed by 26 weeks of personalized smart-messaging (relapse prevention planning [RPP]) or receive treatment as usual (TAU). Outcomes were assessed at discharge and 6- and 12-month follow-up using the Patient Health Questionnaire-9 (PHQ-9; primary outcome), Generalized Anxiety Disorder-7 (GAD-7), and Work and Social Adjustment Scale (WSAS). Linear mixed-effects models examined differences between groups over time while controlling for pre-treatment severity. Feasibility and acceptability were assessed through recruitment, retention, engagement, and withdrawal rates. RESULTS Of 84 eligible patients approached, 57 (68%) consented to participate and 29 (56%) completed the final follow-up assessment. Engagement with smart-messaging was high: 92% of participants responded to at least one weekly prompt and there were no withdrawals from the messaging intervention. For depression, the RPP group demonstrated significantly lower PHQ-9 scores across follow-up than TAU (adjusted mean 6.08, SE 0.95 vs 10.46, SE 0.97; F1,44.9=9.90, P=.003). Significant between-group differences were observed at 6 months (mean difference 5.92, 95% CI 2.40-9.44; P=.001; d=1.08) and 12 months (mean difference 5.41, 95% CI 1.76-9.07; P=.005; d=1.06). The group × time interaction was not statistically significant (F2,36.3=2.73, P=.079). For anxiety, the RPP group reported lower GAD-7 scores than TAU (mean difference 3.80, 95% CI 1.35-6.25; F1,49.8=9.73, P=.003), although the interaction effect was not significant (F2,37.4=0.04, P=.965). Functional impairment showed a similar pattern, with a significant advantage for RPP at 6 months (mean difference 9.09; P=.029; d=1.16), but the overall group effect did not reach statistical significance (F1,20.0=4.08, P=.057). CONCLUSIONS Personalized relapse prevention planning supported by smart-messaging was feasible, highly acceptable, and associated with sustained improvements in depression and anxiety following psychological therapy for cancer patients. Although findings are promising, the non-randomized design limits causal inference. A fully powered randomized controlled trial is warranted.