Background: Walking has proven to be beneficial for cognition in healthy sedentary older people. The aim of this study was to examine the effects of a walking intervention on cognition in older people with dementia. Methods: 97 older nursing home residents with moderate dementia (mean age 85.4 years; 79 female participants; mean Mini-Mental State Examination 17.7) were randomly allocated to the experimental or control condition. Participants assigned to the experimental condition walked for 30 min, 5 days a week, for 6 weeks. To control for personal communication, another group received social visits in the same frequency. Neuropsychological tests were assessed at baseline, directly after the 6 week intervention and again 6 weeks later. Apolipoprotein E (ApoE) genotype was determined. Results: Differences in cognition between both groups the three assessments were calculated using a linear mixed model. Outcome measures included performance on tests that formed three domains: a memory domain, an executive function domain and a total cognition domain. Results indicate that there were no significant time x group interaction effects or any time x group x ApoE4 interaction effects. Conclusion: Possible explanations for the lack of a beneficial effect of the walking programme on cognition could be the level of physical activation of the intervention or the high frequency of comorbid cardiovascular disease in the present population of older people with dementia. Trial registration number: Nederlands Trial Register, TC1230, http://www.trialregister.nl/trialreg/index.asp.