There are limited data on long-term outcomes of mothers or their offspring following exercise interventions during pregnancy. We assessed long-term effects of an exercise intervention (home-based stationary cycling) between 20-36 weeks of gestation on anthropometry and body composition in mothers and offspring after 1 and 7 years. 84 women were randomised to intervention or usual activity, with follow-up data available for 61 mother-child pairs (38 exercisers) at 1 year and 57 (33 exercisers) at 7 years. At 1 year, there were no observed differences in measured outcomes between mothers and offspring in the two groups. At the 7-year follow-up, mothers were mostly similar, except that exercisers had lower systolic blood pressure (-6.2 mmHg; p = 0.049). However, offspring of mothers who exercised during pregnancy had increased total body fat (+3.2%; p = 0.034) and greater abdominal (+4.1% android fat; p = 0.040) and gynoid (+3.5% gynoid fat; p = 0.042) adiposity compared with controls. Exercise interventions beginning during pregnancy may be beneficial to long-term maternal health. However, the initiation of exercise during pregnancy amongst sedentary mothers may be associated with adverse effects in the offspring during childhood. Larger follow-up studies are required to investigate long-term effects of exercise in pregnancy.