PURPOSE: As shared family context may be an important influence on mental health, and gender differences in mental health, in later life we investigated how gender, family-related variables and gender roles were associated with mental health in older married couples. METHODS: Using data on a sample of 2,511 married couples born between 1923 and 1953 (drawn from the British Household Panel Survey) we analysed differences in the mental health of husbands and wives by fertility history, length of marriage, presence of co-resident children, reported social support, hours of household work, attitudes to gender roles and health of husband and wife. Mental health in 2001 was measured using the General Health Questionnaire (GHQ-12). Multilevel modelling was used to assess effects in husbands and wives and variations between husbands and wives. RESULTS: Results showed that although the mental health of married couples was correlated, wives had poorer mental health than their husbands. The gender difference was smaller in couples who lived with a child aged 16 or more (and had no younger co-resident children) and in couples in which both spouses had experienced early parenthood. The influence of individual and family characteristics on mental health also differed between husbands and wives. For husbands, early fatherhood and co-residence with a child or children aged 16 or more increased the odds of poor mental health. For wives, having had a child when aged 35 or more appeared protective while having traditional gender role attitudes increased the odds of poorer mental health. CONCLUSIONS: The role of family characteristics in the shared marital context has complex associations with mental health, some of which seem gender specific. Although wives express more mental distress, husbands in general show poorer mental health related to family characteristics.