The aim was to test if targeting managers with an educational intervention reduces absence among pregnant employees. The study was a non-blinded cluster randomised trial conducted in hospitals and daycare institutions from two administrative Danish Regions and two Danish municipalities. Clusters (work units) were assigned randomly and non-blinded to either (1) intervention, where all managers were invited to participate in a 3-hour seminar addressing needs and options for adjustment of work in pregnancy, or (2) control, with practice as usual. The primary outcome based on payroll data was long-term pregnancy-related absence, defined as ≥12.5% cumulated absence during pregnancy weeks 1–32. Intention-to-treat analysis was applied using mixed logistic regression. Ninety work units were included (56 hospital departments and 34 daycare units) with 451 pregnant employees in the intervention group and 464 in the control group. Work units had on average 11 pregnant employees with no difference between the groups. 103 of the 216 invited managers (48%) participated in a the 3-hour seminar. In the intervention group, 154 (34%) had long-term pregnancy-related absence during pregnancy weeks 1–32 vs 166 (36%) in the control group. Relative odds of having long-term pregnancy-related absence, when being in the intervention group, was 1.06 (95% CI 0.71 to 1.58), with an interclass correlation coefficient of 0.07. An educational intervention targeting managers did not reduce pregnancy-related absence among pregnant employees. NCT03002987.