Background: We examined indicators of job strain in relation to level and change in episodic memory in the years leading up to as well as following retirement. Methods Our analyses centre on 3779 individuals from the nationally representative Health and Retirement Study (baseline age 57.3 years) who reported gainful employment in an occupation for 10+ years prior to retirement, and who were assessed for episodic memory performance over up to 20 years (median 8 waves over 16 years). We used ratings from the Occupational Information Network (O*Net) to score occupations for job control and job demands, and to measure job strain (job demands/job control). Results: Controlling for sociodemographic characteristics, depressive symptoms, and cardiovascular disease, less job control and greater job strain were not significantly associated with change in episodic memory in the period leading up to retirement, but were associated with significantly poorer episodic memory at retirement and an accelerated rate of decline in episodic memory following retirement. The results did not vary for men and women or by self-employment status. Conclusions: Job strain expressed mainly as low job control is linked to poorer episodic memory at retirement and more decline after retirement. Job characteristics appear to have implications for cognitive ageing independent of relevant confounds.