This study investigated relations between general cognitive ability (g) and aspects of time perspective, i.e. habitual ways of relating to the past, present, and future, in a sample of older adults (60-90 years, N = 438). In main focus was a measure of deviations from a balanced time perspective (DBTP), reflecting the differences between proposed ideal and observed score profile on the Zimbardo Time Perspective Inventory (S-ZTPI). A current g factor reflecting four cognitive markers was negatively related to DBTP (beta = -0.31), with a higher estimate (beta = -0.40) for a latent DBTP factor. For a subset of male participants (n = 129), cognitive test score from age 18 were retrieved. In that sample the g factor in youth predicted DBTP scores obtained around 52 years later (beta = -0.31, p < .01) nearly as well as current g (beta = -0.39). In line with prior studies, the Present Fatalistic dimension was a main source of the covariation of g and DBTP, but deviation scores for each of the three temporal frames (past, present, future) were significantly associated with g as well. Variations in recent stress did not account for these associations. Multi-group latent level analyses revealed a magnified g-DBTP association in old-old age (beta = -0.57 and beta = -0.81 in the old-old group for a latent DBTP factor), with a similar pattern for Present Fatalistic and Past Negative. Together, the results demonstrate a substantial association between g and time perspective in late adulthood, a relationship that may have been established early as judged from a relation to the age 18 g factor. A magnified association in in old-old age might reflect a more noticeable impact of age-related cognitive deficits on everyday functioning and thereby aspects of time perspective (e.g. increase present fatalism). Impairments in cognitive processes that allow for a flexible shift between temporal frames could also be factor, something which needs to be evaluated in future studies.