To characterize a rabbit model of Mycobacterium chelonae keratitis after lamellar keratectomy and assess the effectiveness of fluoroquinolone therapy. University Laboratory, University of California, Irvine, California, USA. Twenty-eight New Zealand white rabbits had unilateral lamellar keratectomy with placement of 2.5 x 10(5) colony-forming units of log-phase M chelonae under each flap. Eyes (7 per group) were randomized and treated with sterile balanced salt solution, gatifloxacin 0.3%, ciprofloxacin 0.3%, or levofloxacin 0.5% 4 times daily. Two masked observers examined all eyes on days 2, 5, and 7 and weekly for 4 weeks. Severity of disease and bacterial culture results were the main outcomes measured. The means and standard deviations were calculated, and differences between the groups were statistically analyzed. All eyes developed clinical disease. At the time the rabbits were killed, eyes treated with balanced salt solution, ciprofloxacin, levofloxacin, and gatifloxacin were culture positive in 6 (85.7%), 7 (100%), 6 (85.7%), and 3 (42.9%) of 7 eyes per group, respectively. Frequency of positive culture and the severity of clinical disease in gatifloxacin-treated eyes were significantly less (P < .05) than in the other groups combined. The rabbit model of M chelonae keratitis was successfully developed in our study. A fourth-generation quinolone (gatifloxacin) showed the best performance among the fluoroquinolones tested in our experimental approach. The fourth-generation fluoroquinolone, gatifloxacin, could be effectively used for the treatment of mycobacterial keratitis.