The association of socioeconomic factors, health-related factors, and social support with depressive symptoms has been extensively studied. However, most epidemiological studies have focused on a few factors such as marital status, social class, and employment. In this study of middle-aged women we analyzed both univariate and multivariate associations of socioeconomic factors, perceived physical health factors, and social support with self-rated depressive symptoms measured with the Beck Depression Inventory. A nationwide sample (n = 1851) of Finnish women aged 48-50 years was analyzed. Socioeconomic, health-related, and social support factors were all measured with single items. All variables, except level of urbanization, were significantly associated with depressive symptoms in univariate analyses. Multivariate associations were examined with standard multiple regression analyses in three stages: first with the socioeconomic factors, then with socioeconomic and health factors, and finally entering the social support factors into the regression model. In the first model, being either on a sick leave or at home as working status, being divorced, and having low income reached significance. In the second model, being divorced and having poor perceived physical health were the most powerful among the five significant variables. In the last model, explaining 32% of the variance, the most powerful predictors of depressive symptoms were perceived current physical health, satisfaction with received social support, and quality of intimate relationships. For the prevention or intervention of depressive symptoms among middle-aged women in the population subjects with concurrent subjective or objective health problems and poor social support seem to comprise a particularly important target group.