Peer support has been used effectively in a variety of patient populations, but its effectiveness in improving outcomes in persons with chronic heart failure has not been explored. We trained 9 persons with heart failure to mentor other heart failure patients and tested the effectiveness of this approach in a randomized controlled clinical trial. A low proportion (37%) of the eligible population of hospitalized patients agreed to participate. At the end of the 3-month trial, there was significantly higher heart failure self-care in the intervention group (P < .05). The only difference in social support was a significant decline in perceived support reciprocity in the intervention group (F = 5.94, P = .004). No significant group differences in heart failure readmissions, length of stay, or cost were evident at 90-days, although the heart failure readmission rate was 96% higher in the intervention group when compared to that in the control group. The reasons for low overall enrollment and high readmission rates in the intervention group require further study. Including additional self-care education by a professional, rather than leaving all the education to the mentor, could strengthen the peer support intervention trialed in this study. Small group meetings may be less intrusive and more desirable for this patient population.