We evaluated the ability of enterobacterial repetitive intergenic consensus sequence-based PCR (ERIC-PCR) to differentiate 95 Streptococcus pyogenes strains with M or T serotypes isolated from sporadic streptococcal infections as compared with M or T serotypings and pulsed-field gel electrophoresis (PFGE). Although the ERIC-PCR had less discriminatory power, defined as the ability to divide the strains with the same serotypes into the different sub-types, than PFGE, it consistently classified the strains into 16 patterns with a high correlation with M or T serotyping. The PCR method further discriminated 4 M or T serotypes into sub-types. The application of ERIC-PCR to 5 outbreaks of streptococcal infection produced the results that agreed closely with those of T serotyping and PFGE. ERIC-PCR has sufficient discriminatory power and is a quick and relatively easy technique, making it useful for routine epidemiological investigations.