A series of four agreement studies (classification sessions) were conducted to support the development and validation of a comprehensive pediatric long bone fracture classification system. This system follows the principle of the Müller-AO classification for long bones in adults and integrates most relevant existing pediatric classification systems. The diagnosis includes the distinction between epiphyseal (E), metaphyseal (M), or diaphyseal (D) fractures, as well as identification of child-specific features. This article describes the proposed system in some detail. Digital standard preoperative anteroposterior and lateral radiographs from 267 consecutive pediatric patients (<16 years old and open physis) with single fractures of the distal humerus, radius, or tibia were collected at a single university children's hospital. Fractures were classified independently by five experienced pediatric surgeons. The classification process was assessed for reliability using the kappa coefficient and accuracy using latent class modeling separately for each bone for bone type, and separately for each bone type for child codes. At the last classification session, kappa values for E-M-D and child code classifications were mostly above 0.90, and accuracy estimates were between 75% and 100% for different surgeons, types, and bones. Disagreement and misclassification of fractures were overall very low; hence, experienced and trained surgeons can classify pediatric long bone fractures using the proposed system with high accuracy based on standard radiographic views. The authors encourage wide consultation and further evaluation of this proposed pediatric long bone classification system with a larger number of future users with different training before being used for documentation and clinical studies.