Background <p>High-stakes examinations represent a significant but underrecognized threat to adolescent mental health, contributing to elevated symptoms of depression, anxiety, and insomnia. Despite the global scale of this problem, effective interventions during these critical periods remain scarce, largely due to implementation barriers in both clinical and educational settings. Scalable, low-resource solutions are urgently needed to address this mental health gap in adolescent care.</p> Methods <p>In a cluster randomized controlled trial (ChiCTR2200058881, <i>N</i> = 587), we examined the efficacy of the Guided Narrative Technique (GNT), a brief digital writing intervention, compared to a neutral writing group (NWG). Adolescents preparing for China’s College Entrance Examination within 100 days (<i>M</i>_age = 18.23, <i>SD</i>_age = 0.60) were assigned to GNT (<i>n</i> = 290) or NWG (<i>n</i> = 297) through class-level cluster randomization and completed three consecutive 20-minute daily sessions. The primary outcome was test anxiety, assessed across the intervention and follow-up period. Secondary outcomes were depression, general anxiety, and insomnia, assessed at baseline, post-intervention, and 15-day follow-up.</p> Results <p>For the primary outcome, GNT did not produce significantly greater reductions in overall test anxiety than NWG in the full sample. However, GNT was associated with greater reductions on the TAI worry subscale, representing the cognitive component of test anxiety (<i>d</i> = 0.18, 95% CI [0.01, 0.36]) in exploratory subgroup analyses among adolescents with elevated baseline test anxiety. For secondary outcomes, compared with NWG, GNT resulted in significantly greater reductions in depression (<i>d</i> = 0.35, 95% CI [0.16, 0.54]), general anxiety (<i>d</i> = 0.37, 95% CI [0.18, 0.56]), and insomnia (<i>d</i> = 0.23, 95% CI [0.04, 0.42]) during the intervention, with between-group differences also observed for depression and general anxiety at follow-up.</p> Conclusions <p>GNT did not significantly reduce overall test anxiety, but showed preliminary benefits for depression, general anxiety, insomnia, and the worry component among adolescents with high baseline anxiety, warranting further evaluation in adequately powered trials.</p> Trial registration <p>The study was registered as ChiCTR2200058881.</p>