At present, no standard protocol has been described to detect the presence of Escherichia coli O157 in cattle faeces. Therefore, the sensitivity of 26 different isolation methods was determined in order to recommend a method of choice. Faeces samples from 17 different beef cattle at a farm previously found positive for E. coli O157 were subdivided into a total of 40 samples. It was not known whether the 17 cattle shed E. coli O157 at the time of sampling. At another farm where cattle have been found negative for E. coli O157 on different occasions, five faeces samples were collected. Two methods yielded the highest sensitivity (74%): 6h enrichment in modified tryptone soya broth supplemented with novobiocin (mTSBn) followed by an immunomagnetic separation (IMS) with (i) Dynal beads or (ii) Captivate beads and selective plating on Rainbow agar (RA) plates. Enrichment for 6h was significantly better than 24h enrichment. Only after 24h, buffered peptone water (BPw) was significantly better than mTSBn. A sensitivity of 82% was obtained only when the two most sensitive tests were done simultaneously. Because none of the tests gave 100% sensitivity, it can be concluded that isolation rates of E. coli O157 from bovine faeces using only one of the tested procedures results in an underestimation of the incidence of E. coli O157 in cattle. Performing more than one test on the samples must be considered.