This study examined the relationship between physical health, social support, and depression in a married, middle-aged/older adult sample in which at least one partner has heart disease. The data was obtained from a national longitudinal study the Health and Retirement Survey (HRS) and is composed of selected respondents and their spouses. The HRS Wave 1 data that was used for these analyses was collected in 1992 and 1993. This study tested a stress buffer model predicting the relationship between physical health, social support, and depression. For study inclusion, participants must have been diagnosed with cardiovascular disease and received treatment in the last year. A heart disease construct was developed by calculating the level of disease by the number of conditions and medical treatments received within the last year. A second health category for other chronic health conditions included diabetes, arthritis, cancer, and chronic pain. These constructs were combined into a total disease construct, which provided a broad measure of health problems typical of an older adult population. Social support was determined by respondents' satisfaction with friends, neighbors, family, their marriage, and enjoyable time spent with their spouse. Social support was subdivided into two constructs separating spousal support from social support sources outside the marriage. The Center for Epidemiological Studies Depression short-form (CES-D) calculated depression scores. Findings support a stress-buffering model among older married adults with chronic diseases. Hierarchical multiple regressions found the following main effects predicted Depression: Total Disease (Beta=. 03, p<. 000), Exercise (Beta=-.11, p<. 000), Smoking (Beta=. 04, p<. 001), General Support (Beta=-.21, p<. 000), Spousal Support (Beta=-.19, p<. 000). The Total Diseases by Spousal Support interaction was a significant predictor of Depression for men and women (Beta= -.04, p<. 000) and Total Disease by Spousal Support was also a significant predictor for men and women (Beta=-.03, plt;. 000). For men with Heart Disease, Total Disease by Spousal Support was a stronger predictor (Beta=-.03) than it was for women with Heart Disease (Beta=-.10). These results may partially explain gender differences in heart disease patients and suggests several psychological interventions that could be beneficial.