BACKGROUND: Essential tremor has been associated with cognitive impairment. Mild action tremor labelled physiological tremor is present in all humans. Tremor increases with age. It is unknown if mild action tremor is associated with cognitive performance in older people. OBJECTIVE: To test the hypothesis that mild-mostly non-pathological-action tremor is associated with reduced cognitive performance in community-dwelling older adults. METHODS: The Cooperative Health Research in the Region of Augsburg (KORA)-Age study assessed health-related parameters including cognition in individuals between 65 and 93 years of age. We measured tremor using semi-quantitative ratings of Archimedes-spiral drawings. Participants taking Parkinson medication, with a self-reported neurologic disease or with a spiral rating > 5 indicating certainly pathologic tremor were excluded. The modified Telephone Interview for Cognitive Status (TICS-m) was used to measure cognition. We analyzed the association between tremor and cognition using multivariable adjusted regression models in the whole population and after exclusion of participants with a spiral rating > 3. We employed stepwise backward variable selection to build a final regression model. RESULTS: Tremor severity was associated with cognitive performance. Age, sex, Parkinson score, frailty, tremor score, physical activity, sleep-duration and previous stroke entered the final statistical model. Judged by the differences of adjusted r-squared between models, age was the most important predictor of cognition followed by sex, Parkinson score, frailty, and tremor severity. Sensitivity analyses excluding participants with likely pathological tremors did not attenuate the association between tremor severity and cognitive performance. CONCLUSION: Even mild action tremor severity is associated with reduced cognitive performance in older people. We hypothesize that aging-related brain changes are a common cause of cognitive impairment and tremor.