AIM: To evaluate the accuracy of eight different intraocular lens (IOL) power calculation formulas for a segmented multifocal IOL. METHODS: A total of 53 eyes of 41 adult cataract patients who underwent phacoemulsification and implantation with the SBL-3 segmented multifocal IOL between January 1, 2017 and January 31, 2019 were included in this retrospective study. Preoperative biometry measurements were obtained using an IOL Master. Manifest refraction was performed at least 4wk postoperatively. Accuracy of the eight formulas [Barrett Universal II, Emmetropia Verifying Optical (EVO), Haigis, Hill-RBF 2.0, Hoffer Q, Holladay 1, Kane, and SRK/T] was analyzed. RESULTS: Using current lens constants, all formulas exhibited errors of slight myopic shift in refractive prediction. The Barrett Universal II formula had a significantly lower median absolute error (MedAE) than did Holladay 1 (P=0.02), Kane (P=0.001) and Hill-RBF 2.0 (P<0.001) formulas. The Haigis formula had a lower MedAE value than did the Hill-RBF 2.0 formula (P=0.005). Differences in MedAE values among SRK/T, EVO and Hoffer Q formulas were not significant. After optimizing lens constants, the MedAE values of all formulas were reduced; significant changes were noted for EVO (P=0.022), Haigis (P=0.048), Hill-RBF 2.0 (P=0.014), Holladay 1 (P=0.045) and Kane (P=0.022) formulas. All formulas performed equally well after optimization of lens constants (P=0.203). CONCLUSION: All eight formulas tend to result in a myopic shift when using current lens constants. Optimized lens constants improve the accuracy of these formulas among adult Chinese patients.