Previous studies have reported that dissociative symptoms (DIS) are associated with self-harm (SH) in adolescents. However, most of these studies were cross-sectional, which limits the understanding of their theoretical relationship. We aimed to investigate the longitudinal relationship between DIS and SH in the general adolescent population. We used data from the Tokyo Teen Cohort study (N = 3007). DIS and SH were assessed at times 1 and 2 (T1 and T2) (12 years of age and 14 years of age, respectively). DIS were assessed using the parent-report Child Behavior Checklist (CBCL), and severe dissociative symptoms (SDIS) were defined as a score above the top 10th percentile. The experience of SH within 1 year was assessed by a self-report questionnaire. The longitudinal relationship between DIS and SH was examined using regression analyses. Using logistic regression analyses, we further investigated the risk for SH at T2 due to persistent SDIS and vice versa. DIS at T1 tended to predict SH at T2 (odds ratio (OR) 1.11, 95% CI 0.99 to 1.25, p = 0.08), while SH at T1 did not predict DIS at T2 (B = − 0.03, 95% CI − 0.26 to 0.20, p = 0.81). Compared with adolescents without SDIS, those with persistent SDIS had an increased risk of SH at T2 (OR 2.61, 95% CI 1.28 to 5.33, p = 0.01). DIS tended to predict future SH, but SH did not predict future DIS. DIS may be a target to prevent SH in adolescents. Intensive attention should be given to adolescents with SDIS due to their increased risk of SH.