Up to 80 % of the Western female population experience premenstrual syndrome (PMS). Long-term pharmacological therapy is unacceptable to most women, and is not warranted for moderate symptoms. Nutritional therapies are popular, but lack a clear evidence base. Anecdotal evidence suggests beneficial effects of soy isoflavones because of their influence on endogenous oestrogen and actions on specific tissues. The effect of isolated soya protein (ISP) containing 68 mg/d (aglycone equivalents) soy isoflavones (IF) on premenstrual symptom severity was studied in a seven-menstrual cycle, double-blind, placebo-controlled, crossover intervention study in twenty-three women with prospectively confirmed PMS aged 18–35 years and BMI 19–30 kg/m2. ISP containing IF or milk protein placebo was consumed for two complete menstrual cycles. ISP containing IF (genistein, daidzein, equol) were measured in 24 h urine samples. After two cycles of ISP containing IF intervention, total symptoms (F(2,36) 8·20, P=0·000) and physical symptoms (F(2,36) 8·18, P=0·000) were significantly reduced compared with baseline after both active and placebo treatments, although differences between active and placebo treatment were non-significant. Specific premenstrual symptoms, headache (F(2,32) 4·10, P=0·026) and breast tenderness (F(2,32) 4·59, P=0·018), were reduced from baseline after soy IF, but not milk protein placebo. Cramps (F(2,32) 4·15, P=0·025) and swelling (F(2,32) 4·64, P=0·017) were significantly lower after active treatment compared with placebo. Concentrations of genistein and daidzein were increased following soy IF consumption, but equol production did not enhance symptom reduction. The present study showed that ISP containing IF may have potential to reduce specific premenstrual symptoms via non-classical actions.