Purpose <p>The purpose of this study was to model the effects of variation in the refractive index of the crystalline lens on&#xa0;theoretical biometric measurements.</p> Methods <p>Measurement errors of the axial length of the eye (AL) and lens thickness (LT) were modelled while the lens refractive index and related ocular segment sizes were varied from a 30-year-old model eye. Two approaches of optical path-length (OPL) conversion, the weighted mean ocular refractive index (WMORI) and the segmented refractive index (SRI), were specifically examined.</p> Results <p>In the WMORI approach, when lens refractive index, LT and vitreous chamber depth (VCD) were varied together, AL measurement was either underestimated by 0.209 mm or overestimated by 0.390 mm. The VCD was a major influencing factor. In the SRI approach, the AL measurement error was reduced to 0.077 mm. The same amount of LT measurement error was also found as&#xa0;in the AL.</p> Conclusion <p>The SRI approach provides an improvement over the WMORI in AL measurement because the influence of the VCD was eliminated. Use of a model eye parameter for OPL conversion always causes biometric measurement errors in individual eyes due to the variations in lens refractive index. The newly developed in vivo measurement of lens refractive index needs to be applied in optical biometry to improve AL and LT measurements for individual eyes in the areas of myopia research and cataract surgery.</p>