This was an exploratory investigation of psychosocial risk factors for mortality in women with premature acute myocardial infarction (AMI). Subjects were 83 female participants in the Recurrent Coronary Prevention Project, who were between the ages of 30 and 63 in 1978, nonsmoking, nondiabetic, and at least 6 months beyond their index AMI. Follow-up ranged from 8 to 10 years, with an average of 8.5 years. Six deaths occurred in the 83 women over the follow-up. Univariate predictors of these deaths were arrhythmias on ECG (RR = 7.83, p = .003), being divorced (RR = 6.9, p = .003), being employed without a college degree (RR = 6.8, p = .03), and the inverse of Type A behavior, time urgency, and emotional arousability (p = .03; .005; .006, respectively). Multivariate stepwise logistic regression analysis produced a solution that included as independent predictors: arrhythmias on ECG (RR = 4.01, p = .004), being divorced (RR = 3.43, p = .01), and the inverse of time urgency (RR = 0.35, p = .02). In the multivariate model, "divorced" was interchangeable with "employed without a college degree" and "time urgency" was interchangeable with "emotional arousability." This small sample precludes firm conclusions, but provides a basis for hypothesis development.