Circadian rhythms and sleep are essential for physical health and mental wellbeing as disrupted circadian rhythms have been associated with several diseases. Therefore, understanding the development of the circadian rhythm and its changes throughout later-life is an important step towards improving public health. Ageing is associated with sleep disruption, but the relationship between ageing and sleep has traditionally been assessed using cross-sectional studies that ignore changes within individuals as they age. In addition, early-life light is known to influence the development of circadian rhythms in adulthood, but the mechanism linking perinatal light experiences to later-life circadian rhythms is unknown. Using the UK Biobank (N=502,536) and the University of Manchester Longitudinal Study of Cognition in Normal Healthy Old Age (N=6,375) cohorts, we investigated longitudinal sleep trajectories during ageing and early-life seasonal programming of circadian rhythms in adulthood. We found in the longitudinal mixed models that older adults have decreased sleep efficiency and early sleep time. Belonging to the high sleep efficiency latent class, which has a minimal decline in sleep efficiency, was associated with a lower prevalence of hypertension, circulatory problems, arthritis, breathing problems and recurrent depression than the low efficiency latent class. Results showed a higher risk of hypertension, metabolic syndrome and mortality in the evening-type latent class compared to morning-type individuals. Finally, we identified 67 suggestively significant genomic loci that have genotype-ageing interaction to predict chronotype trajectory, with the strongest associated gene being ALKBH5, which has functions of DNA repair and epitranscriptomic regulation. We showed that birth rate, basal metabolic rate, general health, reaction speed and sleep are seasonally rhythmic in human. We showed that season of birth was associated with basal metabolic rate, reaction speed, allergy, insomnia and chronotype in adulthood. In addition to perinatal season, breastfeeding appeared to influence later-life circadian rhythms and sleep. We also identified 25 genomic loci that may have genotype-season of birth interaction in determining chronotype. A potential candidate gene for postnatal seasonal programming was SIRT1, which has previously been implicated in postnatal programming, ageing and longevity. Lastly, epigenome-wide association analysis of human hypothalamus and inferior temporal gyrus showed that season of birth may result in persistent DNA methylation that is associated with circadian rhythms in adulthood. Overall, circadian rhythms are influenced by ageing and seasonal effects. Genetic diversity may explain the heterogeneity in ageing-related change of sleep timing and postnatal environmental programming of later-life chronotype. Finally, season of birth may imprint later-life circadian rhythm via epigenetic modifications. Understanding the genetic and epigenetic mechanisms behind ageing and early-life programming of circadian rhythm is a key first step to preventing, diagnosing and treatment of future health problems in predisposed individuals.