A social gradient in coronary heart disease (CHD) has been documented in a variety of settings, predominantly among men. This study aimed to establish whether a social gradient in CHD existed in a group of Swedish women and whether it could be explained by established coronary risk factors or psychosocial factors. The Women's Lifestyle and Health Cohort Study includes 49,259 women from Sweden aged 30-50 years at baseline (1991-1992), when an extensive questionnaire was completed. There was complete follow-up through linkages to national registries until the end of 2002, during which time 210 cases of incident fatal CHD or nonfatal myocardial infarction occurred. Risk of CHD was significantly inversely related to years of education, the socioeconomic status proxy (hazard ratio comparing the lowest with the highest education group = 3.3, 95% confidence interval: 2.2, 4.7). This association was reduced after adjustment for established coronary risk factors (smoking, body mass index, alcohol consumption, diabetes, hypertension, exercise; hazard ratio = 1.9, 95% confidence interval: 1.3, 2.8). Job strain and social support were weakly related to CHD and did not explain the gradient by years of education. Self-rated health was strongly related to CHD, mediated by established coronary risk factors. Results show a strong gradient in CHD by years of education explained by established coronary risk factors but not by job strain or social support.