Purpose: To evaluate the in vivo efficacy of rose bengal (RB)-mediated photodynamic antimicrobial therapy (PDAT) for treatment of Acanthamoeba castellanii keratitis (AK). Materials and Methods: An animal (rabbit) AK model was successfully achieved via intrastromal inoculation of a suspension of A. castellanii cells and trophozoites. Prior to RB-PDAT (pre-treatment, day-5), the severity of the induced corneal infection was graded numerically for epithelial defects, stromal edema, neovascularity, and stromal opacity/infiltration. The right eyes of rabbits (n = 18) were divided equally into three groups (n = 6/group): control (no treatment); 0.1% RB+518 nm irradiation (5.4 J/cm(2)); and 0.2% RB+518 nm irradiation (5.4 J/cm(2)). On post-treatment day-5, animals were euthanized, after which corneal buttons were excised and submitted for real-time polymerase chain reaction (RT-PCR) analysis. Results: Post-treatment clinical scores of the 0.1 and 0.2% RB groups indicated significant improvement compared to control group scores (pre-treatment clinical scores; 5.17 +/- 0.98, 7.50 +/- 0.62, and 6.17 +/- 0.70 and post-treatment clinical scores; 4.50 +/- 0.56, (p = .043), 3.50 +/- 0.99 (p = .039), 6.83 +/- 1.66 (p = .34), respectively). RT-PCR analysis revealed that the mean cycle threshold (Ct) values were significantly higher in treated-group corneas compared to control-group corneas, with no significant differences between treated-groups (Mean Ct values; 34.33, 34.5, and 29.67 for 0.1 and 0.2% RB, and control groups). There was a statistically significant negative correlation between post-treatment clinical scores and Ct values (r = -0.474, p-value 0.047). Conclusions: Our results demonstrate that RB-PDAT is effective in decreasing the parasitic load and clinical severity of AK.