Objectives To compare expulsions and adverse events (AEs) between immediate and delayed insertion of a levonorgestrel‐releasing intrauterine system (LNG‐IUS) following medical termination of pregnancy (MTOP). Design Randomised controlled trial. Setting Helsinki University Hospital, Finland, January 2013–December 2014. Population Cohorts of 102 (gestational age 64–84 days, late first trimester) and 57 (gestational age 85–140 days, second trimester) women requesting MTOP and LNG‐IUS contraception. Methods LNG‐IUS insertion occurred immediately (same day) or 2–4 weeks following MTOP. Follow‐up visits were at 2–4 weeks, 3 months, and 1 year. Main outcome measures LNG‐IUS expulsion by 3 months and 1 year. AEs and bleeding profiles within 3 months. Results Following late first‐trimester MTOP the LNG‐IUS expulsion rates by 3 months were 14 (27.5%) in the immediate‐insertion group and two (4.0%) in the delayed‐insertion group (risk ratio, RR 6.86; 95% confidence interval, 95% CI 1.64–28.66). By 1 year the expulsion rates were 17 (33.3%) and six (12.0%) (RR 2.78, 95% CI 1.19–6.47). Following second‐trimester MTOP LNG‐IUS expulsion rates by 3 months and 1 year were five (18.5%) in the immediate‐insertion group and one (3.6%) in the delayed‐insertion group (RR 5.19, 95% CI 0.65–41.54). No differences in AEs and bleeding profiles emerged between the groups. Conclusions Immediate LNG‐IUS insertion after late first‐ or second‐trimester MTOP is feasible, does not increase the complication rate, or alter the uterine bleeding patterns; however, immediate insertion increased the expulsion rate, which may limit the cost‐effectiveness. Tweetable Immediate insertion of LNG‐IUS following MTOP at 9–20 weeks of gestation is feasible and safe. Tweetable Immediate insertion of LNG‐IUS following MTOP at 9–20 weeks of gestation is feasible and safe.