Objective Cognitive performance shows a marked deterioration in close proximity to death, as postulated by the terminal decline hypothesis. The effect of education on the rate of terminal decline in the oldest people (i.e. persons 85+ years) has been controversial and not entirely understood. In the current study, we investigated the rate of decline prior to death with a special focus on the role of education and socioeconomic position, in two European longitudinal studies of ageing: the Origins of Variance in the Old‐Old: Octogenarian Twins (OCTO‐Twin) and the Newcastle 85+ study. Methods A process‐based approach was used in which individuals' cognitive scores were aligned according to distance to death. In a coordinated analysis, multilevel models were employed to examine associations between different markers of cognitive reserve (education and socioeconomic position) and terminal decline using the mini‐mental state examination (MMSE), controlling for age at baseline, sex, dementia incidence and time to death from the study entry to the time of death within each cohort. Results The current findings suggest that education was positively associated with higher MMSE scores prior to death in the OCTO‐Twin, but not in the Newcastle 85+ study, independent of socioeconomic position and other factors such as baseline age, sex and time to death from the study entry. However, education was not associated with the rate of terminal decline in both of these studies. Conclusions Our results offer only partial support to the cognitive reserve hypothesis and cognitive performance prior to death. © 2015 The Authors International Journal of Geriatric Psychiatry Published by John Wiley & Sons Ltd.