Objective: To evaluate the development of refraction, expressed as spherical equivalents, in prematurely born children during the first 10 years of life. Methods: Retinoscopy in cycloplegia was performed at 6 months, 2.5 years, and 10 years of age in 198 prematurely born children from a previous population-based study on the incidence of retinopathy of prematurity. Spherical equivalents were calculated. Myopia was defined as a spherical equivalent of less than 0 diopters (D), clinically significant myopia at 10 years of age as- 1 D or less, and moderate or high myopia as less than- 3D. Hypermetropia greater than+ 3D was regarded as significant. Results: There were no significant differences during the refractive development between the various subgroups of retinopathy of prematurity. Cryotreated eyes had a wider distribution of refractive errors. A multiple regression analysis revealed that the spherical equivalents at 2.5 years of age predicted clinically significant myopia (≤ - 1 D) at 10 years of age. Conclusions: Retinoscopies at 6 months, 2.5 years, and 10 years of age show a similar course of spherical equivalent refractive development regardless of the stage of retinopathy of prematurity. Refraction at 6 months of age is an unreliable predictor, but the refraction at 2.5 years of age seems to be a better tool for identifying refractive errors that will remain at 10 years of age. Methods: Retinoscopy in cycloplegia was performed at 6 months, 2.5 years, and 10 years of age in 198 prematurely born children from a previous population-based study on the incidence of retinopathy of prematurity. Spherical equivalents were calculated. A Myopia was defined as a spherical equivalent of less than 0 diopters (D), clinically significant myopia at 10 years of age as- 1 D or less , and moderate or high myopia as less than- 3D. Hypermetropia greater than + 3D was as significant. Results: There were no significant differences during the refractive development between the various subgroups of retinopathy of prematurity. Cryotreated eyes had a wider distribution of refractive errors . A multiple regression analysis revealed that the spherical equivalents at 2.5 years of age predicted clinically significant myopia (≤ - 1 D ) at 10 years of age. Conclusions: Retinoscopies at 6 months, 2.5 years, and 10 years of age show a similar course of spherical equivalent refractive development regardless of the stage of retinopathy of prematurity. Refraction at 6 months of age is an unreliable predictor , but the refraction at 2.5 years of age seems to be a better tool for identifying refractive errors that will remain at 10 years of age.