Series of animal feces and meat samples artificially contaminated with strains of Escherichia coli O157 isolated from different sources were tested by both an immunomagnetic separation (IMS)-based method and a PCR method using primers specific for a portion of the rfbE gene of E. coli O157. IMS is laborious and time consuming but ends up with the isolation of the pathogen. PCR is fast and less laborious, but it can only be used for screening purposes, so a further culture step is required to isolate the organism. For both fecal and meat samples, the IMS method was found to be more sensitive than the PCR. Furthermore, the detection efficiency of the PCR was influenced by the origin of the fecal sample and the type of meat. For sheep feces, the efficiency of the PCR appeared to be systematically lower than for cattle feces. And the efficiency of the PCR in detecting E. coli O157 in spiked samples of raw minced beef and dry-fermented sausages was systematically lower than in samples of filet americain. Based on this study, it can be concluded that both for animal feces and meat, IMS can be used more successfully to detect E. coli O157 than PCR, because IMS showed to be more sensitive and the outcome was not influenced by the type of animal feces or meat.