The crosstalk between maternal stress exposure and fetal development may be mediated by epigenetic mechanisms, including DNA methylation (DNAm). To address this matter, we collect 32 cord blood samples from low-income Brazilian pregnant adolescents participants of a pilot randomized clinical intervention study (ClinicalTrials.gov, Identifier: NCT02807818). We hypothesized that the association between the intervention and infant neurodevelopmental outcomes at 12 months of age would be mediated by DNAm. First, we searched genome methylation differences between cases and controls using different approaches, as well as differences in age acceleration (AA), represented by the difference of methylation age and birth age. According to an adjusted p-value ≤ 0.05 we identified 3090 differentially methylated positions- CpG sites (DMPs), 21 differentially methylated regions (DMRs) and one comethylated module weakly preserved between groups. The intervention group presented a smaller AA compared to the control group (p = 0.025). A logistic regression controlled by sex and with gestational age indicated a coefficient of −0.35 towards intervention group (p = 0.016) considering AA. A higher cognitive domain score from Bayley III scale was observed in the intervention group at 12 months of age. Then, we performed a potential causal mediation analysis selecting only DMPs highly associated with the cognitive domain (adj. R2 > 0.4), DMRs and CpGs of hub genes from the weakly preserved comethylated module and epigenetic clock as raw values. DMPs in STXBP6, and PF4 DMR, mediated the association between the maternal intervention and the cognitive domain at 12 months of age. In conclusion, DNAm in different sites and regions mediated the association between intervention and cognitive outcome.