Purpose: To evaluate the influence of analysis mode selection on prediction accuracy of corneal astigmatism using Pentacam.Methods: Fifty-nine eyes of 59 patients implanted with toric intraocular lenses (IOLs) were included in the retrospective study. Preoperative corneal astigmatism (total refractive power) measured with Pentacam was analyzed based on 2-, 3-, 4-, or 5-mm ring or zone mode either centered on corneal apex or pupil center. Actual corneal astigmatism was calculated based on residual astigmatism on the corneal plane, surgical-induced astigmatism, and effective toric power on the corneal plane. Prediction error, the difference between actual corneal astigmatism and measured astigmatism, was compared among different analysis modes. Influences of local topography on prediction error were also evaluated.Results: Based on the zone mode, prediction error was lower when centered on corneal apex than on pupil center at different diameters, whereas based on the ring mode, this difference was only seen at 2-mm cornea (all P < 0.05). When centered on the corneal apex, the zone mode showed lower prediction error than the ring mode at 4- and 5-mm corneas (both P < 0.001), regardless of asymmetric or symmetric astigmatism. In symmetric bowtie, the zone mode showed lower prediction error than the ring mode at 2-mm cornea of the small bowtie, and 4- and 5-mm corneas of the large bowtie (all P < 0.05).Conclusions: For toric IOL planning, the corneal apex may be a better reference center. At a cornea diameter ≥4 mm, the zone mode is more accurate than the ring mode. Local topography affects prediction accuracy in the symmetric bowtie.