Background: The effects of prenatal particulate matter with an aerodynamic diameter ranging from 0.1 μm to 2.5 μm (PM2.5) and vitamin D on atopic dermatitis (AD) phenotypes have not been evaluated. DNA methylation and cord blood (CB) vitamin D could represent a plausible link between prenatal PM2.5 exposure and AD in an offspring. Objective: To determine the critical windows of prenatal PM2.5 exposure on the AD phenotypes, if vitamin D modulated these effects, and if placental DNA methylation mediated these effects on AD in offspring. Methods: Mother–child pairs were enrolled from the birth cohort of the Cohort for Childhood Origin of Asthma and allergic diseases (COCOA) study. PM2.5 was estimated by land-use regression models, and CB vitamin D was measured by chemiluminescence immunoassay. AD was identified by the parental report of a physician's diagnosis. We defined the following 4 AD phenotypes according to onset age (by the age of 2 years) and persistence (by the age of 3 years): early-onset transient and persistent, late onset, and never. Logistic regression analysis and Bayesian distributed lag interaction model were used. DNA methylation microarray was analyzed using an Infinium Human Methylation EPIC BeadChip (Illumina, San Diego, California) in placenta. Results: PM2.5 exposure during the first trimester of pregnancy, especially during 6 to 7 weeks of gestation, was associated with early-onset persistent AD. This effect increased in children with low CB vitamin D, especially in those with PM2.5 exposure during 3 to 7 weeks of gestation. AHRR (cg16371648), DPP10 (cg19211931), and HLADRB1 (cg10632894) were hypomethylated in children with AD with high PM2.5 and low CB vitamin D. Conclusion: Higher PM2.5 during the first trimester of pregnancy and low CB vitamin D affected early-onset persistent AD, and the most sensitive window was 6 to 7 weeks of gestation. Placental DNA methylation mediated this effect.