Mental health has decreased in young people since the 1990s, and mental health promotion is an urgent matter. A first step is to identify which social determinants could be of importance for intervention. We used the Stockholm Public Health Cohort, a longitudinal population-based health survey, completed by 31,000 inhabitants in the Stockholm County. We focused on the 18–29 age group, n = 3373 (60% females, 40% males) and aimed at assessing which social determinants predict stable mental health, measured as scoring <3 points on the General Health Questionnaire 12 at all time points: 2002, 2007, 2010, and 2014. Forty-six percent of males and 36% of females reported stable mental health. Among the 17 predictors on sociodemographics, socioeconomics, social capital, health behavior, and victimization, six predicted stable mental health in the following order: occupation and especially employment, emotional support, male gender, being born in Sweden, absence of financial strain, and consumption of fruit and berries. In the 30–84 age group, 66% males and 55% females reported stable mental health. Nine determinants in the following rank predicted stable mental health: absence of financial strain, occupation and especially being self-employed, emotional support, male gender, physical activity, instrumental support, interpersonal trust, community trust, and absence of hazardous alcohol consumption. Interaction analysis showed significant difference between the younger and older group regarding physical activity and absence of financial strain with importance being higher for the older group. Our findings indicate that the determinants of health differ across the life-course with fewer predictors related to social capital and health behavior in the younger group compared to the older. We conclude that health-promoting interventions should be lifespan-sensitive.