Purpose To determine optical biometry data criteria for the detection of abnormal refraction in preschool children, and to evaluate the accuracy of these criteria for detecting amblyopia refractive risk factor (ARF), as defined in the 2013 guidelines of the American Association for Pediatric Ophthalmology and Strabismus (AAPOS). Methods The present study included 200 eyes of 100 preschool children with normal eyes for the experimental determination of criteria and 142 eyes of 71 preschool children for validation of these criteria. Statistical data from normal eyes were used to determine both Bhigh sensitivity failure criterion^and Bhigh specificity failure criterion^associated with corneal astigmatism, interocular difference in axial length, and the prediction interval of a regression formula for predicting corneal power from axial length. Ophthalmological examination of children for validation included testing cycloplegic refraction and optical biometry testing. Outcomes from optical biometry criteria were compared with determination via ophthalmological examination, and the accuracy of the criteria for detecting ARF was evaluated. Results Sensitivity of the Bhigh sensitivity failure criterion^for detecting 2013 AAPOS ARF was 100%, while the specificity was 80.5%. The sensitivity of the Bhigh specificity failure criterion^was 93.3%, while the specificity was 95.1%. Conclusions The criteria derived from optical biometry data in this study exhibited excellent sensitivity and specificity for detecting ARF. This study may lead to a new approach to vision screening in preschool children.