Methicillin-resistant Staphylococcus aureus (MRSA) has become a worldwide problem, although its prevalence varies considerably among countries. The epidemiology of MRSA is now changing; infections are no longer confined to the hospital setting, but also appear in healthy community-dwelling individuals without established risk factors for the acquisition of MRSA. Reported prevalence rates of community-acquired MRSA (CA-MRSA) vary widely among studies—largely because of the different definitions employed and different settings in which the studies have been performed. At present, molecular epidemiological definitions, based on staphylococcal cassette chromosome mec (SCCmec) typing and phylogenetic analyses of the MRSA isolates, are considered the most reliable means by which to distinguish between hospital-acquired MRSA (HA-MRSA) and CA-MRSA. CA-MRSA has been isolated predominantly from skin and soft tissue infections, such as abscesses, cellulitis, folliculitis and impetigo. Although CA-MRSA infections are usually mild, they may also be severe, and can result in hospitalisation and even death. CA-MRSA strains differ from the major pandemic clones of MRSA that account for the majority of epidemic HA-MRSA strains. Differences are found in SCCmec types, bacterial growth rate, and the distribution of antibiotic resistance genes and toxin genes. Mathematical models have shown that CA-MRSA has a high potential to become endemic in the community, and this will impact significantly on the control of MRSA in the hospital setting. Well-designed, community-based studies with adequate risk factor analysis are required to further elucidate the epidemiology of CA-MRSA and to improve strategies to control MRSA in both the community and hospital settings.