Objectives: We studied the predicted value of disease-related characteristics for the ability of children with osteogenesis imperfecta (OI) to walk. Study design: The severity of OI was classified according to Sillence. The parents were asked to report the age at which the child achieved motor milestones, the fracture incidence, and the age and localization of the first surgical intervention. The present main means of mobility was classified according to Bleck. Results: There were 76 replies to the 98 questionnaires, of which 70 were included (type I, 41; type III, 11; type IV, 18). The type of OI was strongly associated with current walking ability, as was the presence of dentinogenesis imperfecta. Patients with type III and IV had a lower chance of ultimately walking compared with those with type I. Children with more than 2 intramedullary rods in the lower extremities had a reduced chance of walking than patients without rods. Rolling over before 8 months, unsupported sitting before 9 months, the ability to get in sitting position without support before 12 months, and the ability to get in a standing position without support before 12 months showed positive odds ratios. In Bleck ≥4, multivariate analysis revealed that only the presence of rodding (yes/no) in the lower extremities had additional predictive value to the type of OI. The presence of dentinogenesis imperfecta and rodding (yes/no) had additional value in Bleck ≥5. Conclusion: The type of OI is the single most important clinical indicator of the ultimate ability to walk. Information about motor development adds little. The early achievement of motor milestones contributes to the ability of independent walking when the type of OI is uncertain. Intramedullary rodding of the lower extremities is primarily related to the severity of the disease and in this way provides consequences for the ability to walk. (J Pediatr 2000;137:397-402)