Dementia affects physical as well as cognitive performance. In individuals with dementia (IWD), decline in physical performance increases with disease progression and is associated with higher functional dependence and decreased quality of life. It is paramount to examine factors that potentially preserve physical performance in IWD, particularly in light of conflicting findings on the effectiveness of physical activity interventions on physical performance of IWD, mainly due to limited number of high-quality studies, large heterogeneity in methods used, or insufficient reporting of methods. The aim of this study was to investigate the effects of a 16-week multimodal exercise program (MEP) combining physical and cognitive tasks on physical performance in IWD, and to identify individual characteristics of MEP responders. A multicenter randomized controlled trial with assessment methods identified by an expert panel was conducted. We included 319 IWD of mild to moderate severity, older than 65 years, who underwent a standardized MEP specifically designed for IWD. At baseline and immediately after the MEP, we assessed physical performance (ie, mobility, balance, and strength) and function of lower extremities (primary outcomes). Potential effects of the MEP on physical performance were identified using 2-factor analyses of variance with repeated measurements within 2 samples (ie, intention-to-treat and per-protocol sample). Additionally, we compared characteristics related to physical performance between positive, non-, and negative responders. Neither analysis procedure revealed statistically significant time×group effects. However, 28% to 40% of participants were positive responders with regard to balance, and strength and function of lower extremities; and these persons had statistically significant lower baseline performance in the corresponding assessments. This randomized controlled trial revealed no overall effects of the MEP on physical performance, probably due to high heterogeneity of the study sample. Findings in responder analysis showed that IWD with lower physical performance at baseline tended to benefit more than those with higher baseline performance. Thus, a higher degree of individualization of the MEP depending on baseline performance on IWD may improve overall MEP effectiveness.