despite the extensive literature on the effectiveness of cognitive training, its effectiveness has not been demonstrated within a population-based long-term care system. To provide cognitive training services to older people with mild dementia, Korea introduced a special dementia rating, as a long-term care grades, in the national long-term care insurance in July 2014. These services are only offered to those with the special dementia rating. This study evaluated the national long-term care insurance-funded cognitive function training programme for older people with mild dementia. data were derived from the Korean National Health Insurance Elderly Cohort database between 2008 and 2015. We compared changes in function between the intervention (n = 352) and control (n = 1952) groups before (2014) and after (2015) introduction of the cognitive function training programme. Difference-in-differences analysis was performed, to compare changes in each score between the intervention and control groups before and after introduction of the cognitive function training programme. introduction of the cognitive function training programme was associated with significantly less cognitive function decline in the intervention group than in the control groups (β = -3.39; standard errors [SE] = 1.14; P = 0.003). A subgroup analysis revealed that this effect increased in subjects in the youngest group, low income bracket, who had a primary caregiver, who were supported in multiple activities of daily living by the primary caregiver, or who were not living alone. introduction of the cognitive function training programme was associated with positive effects on cognitive function.