To interpret changes of balance and mobility in people with Alzheimer's disease (AD), we require measures of balance and mobility that have demonstrated reliability in this population. The aim of the study was to determine the safety, feasibility and retest reliability of clinical and forceplate balance and mobility measurements in people with AD.Relative and absolute reliabilities were examined in 14 older people with mild to moderate AD. Relative reliability was calculated using the intraclass correlation coefficient, two-way mixed model (ICC3,1). Absolute reliability was calculated using the standard error of measurement (SEM), the minimum detectable change (MDC) and the coefficient of variation (CV).All measurements were clinically feasible and could be safely administered. ICC values were excellent and CVs were less than 11% in all clinical balance and mobility measures except the Timed Up & Go test with cognitive or manual task (ICC3,1 = 0.5 and 0.7, and CV = 14% and 10%, respectively). Most balance and mobility measures tested on the Neurocom™ forceplate (modified Clinical Test of Sensory Interaction on Balance, Walk Across (step width, step length parameters), and Sit to Stand (rising index parameter)) had excellent relative reliability (ICC3,1 ranging from 0.75 to 0.91). ICC values were fair to good for the other measures.Retest reliability of the balance and mobility measures used in this study ranged between fair to good, and good to excellent. Clinicians should consider retest reliability when deciding which balance and mobility measures are used to assess people with AD.