Objective To explore the effects of narrative intervention and different intervention models on children at risk for developmental language disorders (DLD). Methods Using a random number table method, thirty children at risk for DLD in Yuhua District, Changsha City, were selected from September to December 2024 and randomly assigned to three groups: individual group (n=10), peer group (n=10), and control group (n=10). The individual and peer groups (2-3 children per peer group) received a 20-hour narrative intervention, while the control group did not receive any intervention. The Little Frog, Where Are You? test was used to assess the children’s narrative abilities, and both global and local coherence error rates were analyzed. Results A total of 24 children were enrolled in the final analysis, including 7 in the individual group, 9 in the peer group, and 8 in the control group. Compared with pre-intervention levels, in the individual group, within-group differences in global and local coherence error rates at 1 week and 3 months post-intervention were statistically significant (all P<0.05) with large effect sizes (Cohen’s |d|>1); in the peer group, only the within-group difference in local coherence error rate at 1 week post-intervention reached statistical significance (P<0.05), with a large effect size (Cohen’s |d|>1); in the control group, within-group differences in global coherence error rate were statistically significant at both 1 week and 3 months post-intervention (both P<0.05), with large effect sizes (Cohen’s |d|>1). Between-group comparisons revealed that at 1 week post-intervention, statistically significant differences in local coherence error rates were observed between both intervention groups and the control group (both P<0.05), with large effect sizes (Cohen’s |d|>1); 3 months post-intervention, the difference in global coherence error rate between the individual group and the control group was statistically significant (P<0.05) and exhibited a large effect size (Cohen’s |d|>1); no statistically significant intergroup differences were identified in any other pairwise comparisons (all P>0.05). Conclusions Early narrative intervention has a preliminary positive effect on enhancing the narrative coherence in children at risk for DLD. In clinical practice, either individualized intervention or peer participatory intervention can be used independently, or the advantages of both approaches can be integrated by alternating individual and peer collaborative interventions.