This study aimed to develop a questionnaire for evaluating total sedentary time (ST) and ST with cognitive activity, and to examine the association between ST and cognitive function among Japanese older adults. The questionnaire to evaluate ST comprised 12 items regarding behavior in specific settings, including 8 items on ST with cognitive activity, in a usual week. Older adults aged ≥75 years who participated in a health check-up assessing cognitive function completed the developed questionnaire and subsequently wore an accelerometer and recorded a diary of ST with cognitive activity for a week as validity measures. Cognitive function was assessed with neuropsychological tests covering 4 domains: memory, attention, executive function, and processing speed. Fifty-two participants were included in the validity analysis. Spearman’s correlation coefficient indicated fair-to-good agreement between the questionnaire-measured and the diary-measured time for ST with cognitive activity (r = 0.59, p < 0.001), but this was not the case for total ST. Bland-Altman plots showed that the questionnaire-measured total ST contained proportional bias (r = 0.51, p < 0.001). Multiple regression analysis (n = 49) showed longer questionnaire-measured ST with cognitive activity was significantly associated with better neuropsychological test scores (attention: β = -0.38, p = 0.025; executive function: β = -0.46, p = 0.003; and processing speed: β = 0.31, p = 0.041), while total ST was not associated with better cognitive performance. The developed questionnaire showed acceptable validity to measure ST with cognitive activity, which was found to be protectively associated with cognitive function.