ABSTRACT: Campylobacter spp. may cause fever, vomiting, and diarrhea in humans. Antibiotic treatment is suggested for patients with severe campylobacteriosis. However, the interpretative criteria for antibiotic susceptibility are inconsistent between Clinical and Laboratory Standards Institute and European Committee on Antimicrobial Susceptibility Testing guidelines. The aim of the study is to investigate the antibiotic susceptibility and prevalence of cytolethal distending toxin genes and to evaluate antibiotic susceptibility testing methods in the clinical laboratories of two tertiary medical centers in Taiwan. In total, 236 bacterial isolates were collected between 2001 and 2014. The disk diffusion and E-test methods were used to evaluate the antibiotic susceptibility, and broth dilution results were used as a reference. The virulence genes cdtA, cdtB, cdtC, and ceuE were detected through polymerase chain reaction. The antimicrobial sensitivity rates for erythromycin, ciprofloxacin, and tetracycline using the broth dilution assay were 80.4, 5.4, and 3.4%, respectively. No significant differences were observed in the antibiotic susceptibility of the isolates obtained from southern and northern Taiwan. However, some differences were observed between species. The susceptibility test for erythromycin (disk diffusion) showed that the isolates with small inhibition zone diameters were all resistant, and five isolates (4.0%) with large IZDs were non-sensitive. The error rate of the disk diffusion method according to the CLSI M45-A3 guideline was 5.4% (8/148). The incompatibility rates between the E-test and broth dilution methods for erythromycin, ciprofloxacin, and tetracycline were 8.0, 5.3, and 1.3%, respectively. The positive rates of the genes cdtA and cdtC were considerably higher in Campylobacter jejuni than in C. coli. Erythromycin is recommended as the first choice of treatment for campylobacteriosis. The disk diffusion method is suitable for prescreening Campylobacter susceptibility by using the CLSI M45-A2 and EUCAST criteria (low error rate of 3.4%). If antibiotic treatment fails or IZDs are between 6 and 20 mm, minimum inhibitory concentration testing by using the E-test method is highly recommended because the results of the E-test and broth dilution methods exhibit high agreement. The error rate of disk diffusion method using CLSI M45-A3 criteria is only slightly higher than B, which is also a suitable criteria.