There is substantial evidence that work plays a significant role in post-retirement health. Yet little is known about its role in when late-life dependency may occur. We examined associations between job strain and the risk of entering late-life dependency. Individually linked nationwide Swedish registers were used to identify people 70+ alive in January 2014, and who did not experience the outcome (late-life dependency) during two months prior to the start of the follow-up. Late-life dependency was operationalized as use of long-term care. Information about job strain was obtained via a job exposure matrice and matched with job titles. Cox regression models with age as time-scale (adjusted for living situation, educational attainment, country of birth, and sex) were conducted to estimate hazard ratios (HR) for entering late-life dependency during the 24 months of follow-up (n=993,595). Having an initial high starting point of job strain followed by an increasing trajectory throughout working life implied a 23% higher risk of entering late-life dependency at a younger age, compared with the reference group (low starting point with a decreasing trajectory). High initial starting point followed by a stable trajectory implied a 12% higher risk of entering late-life dependency at a younger age. High initial starting point followed by a decreasing trajectory implied a 10% risk reduction, and a low starting point with a stable trajectory implied a 22% risk reduction, of entering late-life dependency at a younger age. Reducing stressful jobs across working life may contribute to postponing late-life dependency.