The important relationship between cognition and gait in people with dementia has been explored with dual-task studies using added cognitive tasks. Effects of less commonly studied but also attention-dividing motor dual-tasks are important to assess in this group as they are common in everyday function and may affect gait differently from cognitive dual-tasks. They may also be easier to comprehend allowing their application with more severe cognitive impairment. The aim of this study was to evaluate the effects and feasibility of a motor dual-task (MDT) on gait measures in people with Alzheimer's disease (AD). Thirty people (15 men, mean age ± SD, 80.2 ± 5.8 years) with a diagnosis of probable AD (MMSE range 8-28) walked on an electronic walkway (i) at self-selected comfortable pace and (ii) at self-selected comfortable pace while carrying a tray and glasses. The MDT produced significant decreases in velocity (Baseline=111.5 ± 26.5 cm/s, MDT=96.8 ± 25.7 cm/s, p<0.001) and stride length (Baseline=121.4 ± 21.6 cm, MDT=108.1 ± 21.0 cm, p<0.001) with medium effect sizes, and increased stride time (Baseline=1.11 ± 0.11s, MDT=1.14 ± 0.12s, p=0.001) with small effect size. Measures of spatial (Baseline=3.2 ± 1.0%, MDT=3.9 ± 1.5%, p=0.006) and temporal (Baseline=2.4 ± 0.8%, MDT=2.8 ± 0.8%, p=0.008) variability increased with the motor dual-task, with medium effect sizes. A trend for motor dual-task changes in gait measures to increase with greater disease severity did not reach significance. The tray-carrying task was feasible, even for participants with severe cognitive decline. Further comparison of different types of motor and cognitive dual-tasks may contribute to development of a framework for clinical intervention to improve reduced dual-task walking capacity in people with AD.