Abstract The extent to which long-term individual-oriented flexibility in working hours is associated with working beyond retirement age is not known. The aims of the present study were to identify trajectories of worktime control (WTC) and to examine whether the membership of WTC trajectories was associated with working beyond individual’s pensionable age. A total of 1,953 older employees participated in the study and had data up to 16 years before pensionable age. Group-based latent trajectory modeling was used to identify WTC trajectories and Cox proportional hazard regression models were used to examine the associations of WTC trajectories with duration of employment. Seven trajectories described WTC: “Stable very low” (7%), “Stable low” (21%), “Declined” (12%), “Stable mid-low” (28%), “Improved” (10%), “Stable high” (16%), and “Stable very high” (5%). When compared with the lowest WTC trajectory groups, trajectories of “Stable high/very high” (hazard ratio [HR] 1.34, 95% confidence interval [CI] 1.17–1.54) and “Improved” WTC (HR 1.49, 95% CI 1.25–1.78) were associated with longer duration of employment. Although the memberships of the “Stable high/very high” and “Improved” WTC trajectories correlated with gender, marital status, occupational position, and self-rated health, the association between WTC and duration of employment was not fully confounded or mediated by these factors. These findings support the hypothesis that having improved or constantly high control over working times from midlife to retirement age may prolong working lives at retirement age.