Background Mild cognitive impairment (MCI) is a transition stage between healthy cognition and dementia, and is linked to poorer sleep. Objective, reliable, and low-burden field methods to measure older adult sleep are also currently needed. The MotionWatch8© (MW8) wrist-worn actigraph provides estimates of sleep with 14 days of observation; however, there may be underlying differences in the reliability of sleep estimates based on MCI status. We therefore investigated the number of MW8 monitoring days required to estimate sleep in older adults with MCI and without. Methods Older adults (55+ years; N=151) wore the MW8 for ≥14 days. The Montreal Cognitive Assessment was used to categorize participants with probable MCI (scores of <26/30) and participants without MCI (≥26/30). We calculated intra-class reliability coefficients for 1-, 7-, and 14-days of wear-time, and performed Spearman-Brown predictions to determine the number of monitoring days needed for an ICC=0.80. Results Older adults with MCI were older ( p <0.01), more likely to be male ( p =0.03), and had shorter sleep duration ( p <0.01). Spearman-Brown analyses indicated that the number of monitoring days needed for an ICC=0.80 in older adults with probable MCI was 7 days for sleep duration, 4 days for fragmentation, and 4 days for efficiency; adults without MCI required 4 days for duration, 6 days for fragmentation, and 3 days for efficiency. Conclusions Our results indicate that while the reliability of MW8 estimates of sleep differs based on cognitive status, 7 days of MW8 monitoring provides reliable estimates of sleep for adults with MCI and those without.