BACKGROUND: Marital status is associated with survival. PURPOSE: The aims of this study are to evaluate marital history and timing on mortality during midlife, test the role of pre-marital personality, and quantify the role of health risk behaviors. METHODS: Cox proportional hazard models were run with varying classifications of marital history and sets of covariates. RESULTS: In fully adjusted models compared to the currently married, lifetime marital history predicts premature mortality with never married at 2.33 times risk of death and ever married at 1.64 risk of death. Midlife marital history shows that not having a partner during midlife (hazard ratio (HR) = 3.10 formerly married; HR = 2.59 remaining single) has the highest risk of death. Controlling for personality and health risk behaviors reduces but does not eliminate the impact of marital status. CONCLUSION: Consistency of marital status during midlife suggests that lack of a partner is associated with midlife mortality.