Aim: To investigate the effectiveness of exercise on impairments, function and quality of life (QOL) in people with dementia by undertaking a systematic review and meta-analysis of the literature. Methods: Seven electronic databases (Cochrane, MEDLINE ISI, CINAHL, Web of Science, EMBase, PsycINFO and Scopus) were systematically searched to source for English-language studies published between 1966 and 29 July 2011. Only randomised controlled trials which evaluated the effectiveness of exercise on impairments, function and QOL were considered. Methodological quality of studies was evaluated using the PEDro scale. A standardised data extraction form was used to record relevant information. Studies utilising similar outcomes were pooled by calculating mean or standardised mean differences using the fixed or random effects model depending on study heterogeneity. Studies which were not included in the meta-analysis were reported in a narrative format. Results: Twenty-three of the 381 studies screened, comprising of 1370 participants, were included. The median quality score was 4.5/10 (range 2 to 7). A range of interventions such as walking, strength training and balance programmes were reported. The meta-analyses demonstrated significant standardised mean differences in favour of exercise compared to non-exercise or no intervention controls for cognition (0.61; 95% CI, 0.33-0.88; P <0.01), balance (1.43; 95% CI, 0.10-2.75; P=0.03) and Activities of Daily Living (ADL) (0.33; 95% CI, 0.09-0.57; P=0.01). Non-significant mean differences in favour of exercise were found for gait speed, endurance and depression. Two studies which investigated the effects of exercise on QOL reported contradictory findings. Conclusion: There is evidence to support the effectiveness of exercise on cognition, balance and ADL in people with dementia. The evidence is less conclusive for the effects of exercise on walking ability, depression and QOL. Future high quality studies comprising of large sample sizes with long-term follow-up are required.