The distinction of antimicrobial resistance of clinical bacteria isolated from county hospitals and a teaching hospital was investigated. Disc diffusion test was used to study the antimicrobial resistance of isolates collected from county hospitals and a teaching hospital. The data was analyzed by WHONET5 and SPSS statistic software. A total of 655 strains and 1682 strains were collected from county hospitals and a teaching hospital, respectively, in the year of 2003. The top ten pathogens were Coagulase negative staphylococci (CNS), E. coli, Klebsiella spp., S. areus, P. aeruginosa, Enterococcus spp., Enterobacter spp., otherwise Salmonella spp., Proteus spp., Shigella spp. in county hospitals and Streptococcus spp., Acinetobacter spp., X. maltophilia in the teaching hospital. The prevalence of multi-drug resistant bacteria was 5% (4/86) of methicillin-resistant S. areus (MRSA), 12% (16/133) and 15.8% (9/57) of extended-spectrum beta-lactamases producing strains of E. coli and Klebsiella spp., respectively, in county hospitals. All of the three rates were lower than that in the teaching hospital and the difference was statistically significant (P < 0. 01). However, the incidence of methicillin-resistant CNS (MRCNS) reached to 70% (109/156) in the two classes of hospitals. Generally, the antimicrobial resistant rates in the county hospitals were lower than those in the teaching hospital, except the resistant rates of ciprofloxacin, erythromycin, clindamycin, SMZco which were similar in the two classes of hospitals. There were differences between county hospitals and the teaching hospital in the distribution of clinical isolates and prevalence of antimicrobial resistance. It was the basis of rational use of antimicrobial agents to monitor antimicrobial resistance by each hospital.