Background The Coronavirus Disease 2019 (COVID-19) pandemic elevated the risk for mental health problems in pregnant women, thereby increasing the risk for long-term negative consequences for mother and child well-being. There was an immediate need for easily accessible interventions for pregnant women experiencing elevated levels of pandemic related stress. Methods A three-session intervention “Online Communities” (OC) was developed at the beginning of the Dutch lockdown, and implemented by a team of midwives and psychologists specialized in Infant Mental Health. Pretest (N = 34) and posttest (N = 17) measurements of depressive symptoms, worries about COVID-19 and worries in general, and mother-to-infant bonding were administered, as well as a posttest evaluation. Results At pretest, the OC group was compared to two reference groups of pregnant women from an ongoing pregnancy cohort study: a COVID-19 (N = 209) and pre-COVID-19 reference group (N = 297). OC participants had significantly more depressive symptoms than both reference groups, and less positive feelings of bonding than the COVID-19 but not the pre-COVID-19 reference group. Compared to pretest, significant decreases in depressive symptoms (with significantly less participants scoring above cut-off) and worries about COVID-19 (large effect sizes) and worries in general (moderate to large effect size) were found at posttest for the OC participants. No significant improvement was found in bonding. Participants rated the intervention positively. Conclusions The current study provides initial evidence supporting the idea that OC is a promising and readily accessible intervention for pregnant women experiencing stress due to the COVID-19 pandemic, and possibly also applicable to other stressors. Trial registration This intervention was registered in the Netherlands Trial Registration (registration number Trial NL8842, registration date 18/08/2020).