Objective To compare the accuracy of different methods for calculating human lens power when lens thickness is not available and find suitable methods for large-scale studies of refractive development. Methods In this cross-sectional survey， lens power was calculated by three different methods. The three methods used the biometry and refraction data of 378 emmetropic eyes of 189 subjects （age range， 7-14 years）. These three methods consist of the Bennett method， which uses lens thickness， and a modification of the Stenström method and the Bennett-Rabbetts method， both of which do not require knowledge of lens thickness. Lens powers calculated with the modified-Stenström and Bennett-Rabbetts methods were compared for accuracy to those calculated with the Bennett method. Data were analyzed by a paired sign test， Wilcoxon rank sum test and Pearson correlation analysis. Results Using the Gullstrand-Emsley and Bennett-Rabbetts eye models， the modified-Stenström method gave lens powers that were approximately 0.46±0.35 D and 0.29±0.35 D lower than the Bennett lens powers and were significantly different from it （signrank=-159.5， -120， P<0.01）. The Bennett-Rabbetts method gave lens powers that were approximately 0.27±0.35 D and 0.09±0.34 D lower and were significantly different from the Bennett lens powers （signrank=-112.5， -42， P<0.01）. By customizing the c constants， the differences in the two methods were remarkably reduced to nonsignificance （signrank=5， P>0.05）. The largest difference was just 1.35 D. Agreement with the Bennett method was within ±0.50 D for 85.4% of the eyes. The lens power differences determined with the Bennett and Bennett-Rabbetts methods decreased with age for children 7-12 years old and increased with age for children above 12 years old （χ²=314.53， P<0.01）. The power difference between the two methods had a negative correlation with age （r=-0.36， P<0.01）. Conclusion With appropriately customized constants， the Bennett-Rabbetts method provides a good approximation of the Bennett lens power in emmetropic eyes. However， the agreement between the two methods for myopia and hyperopia needs further study.