Aim. This article reports a randomised controlled trial to determine the efficacy of antenatal pelvic floor muscle exercises in the primary prevention of postpartum stress incontinence in primiparous women. Background. Pelvic floor muscle exercises are effective in treating stress incontinence, yet prevention studies demonstrate equivocal findings. Design. Randomised controlled trial. Method. Pregnant women recruited from two hospitals in North-west England were randomised to an intervention (n = 141) or control group (n = 145). Data were collected from 2005–2006. The intervention comprised four sessions of taught pelvic floor muscle exercise training during pregnancy and 8–12 maximal contractions repeated twice daily at home. A modified Bristol Female Lower Urinary Tract Symptom questionnaire, Leicester Impact Scale and Three Day Diary were administered at 20 and 36 weeks of pregnancy and three months postpartum. Results. The intervention group was more likely to exercise their pelvic floor muscles compared to controls at 36 weeks (p = 0·019) and three months (0·022), reporting fewer episodes of incontinence and a lower score on the Leicester Impact Scale. However, these differences were not statistically significant. Conclusion. Significant differences were not demonstrated between the groups in relation to incontinence episodes and degree of bother of symptoms postpartum, although trends indicate a positive effect. Further research is necessary to address issues of adherence and the effect of pelvic floor muscle exercise undertaken during pregnancy on postpartum stress urinary incontinence. Relevance to clinical practice. A proportion of women did not meet the required attendance at antenatal class, furthermore, few exercised their pelvic floor muscles during pregnancy according to instructions. Health professionals need to find ways to instruct and motivate women to perform pelvic floor muscles exercises regularly during pregnancy and the postpartum.