A taxonomy and methodology are developed to index aviation accident models into the case base of an expert system, which utilizes a Case-Based Reasoning approach for reasoning. A group of aircraft accidents selected from five major aviation accident categories is modeled in accordance with the Aviation System Risk Model approach which employs Bayesian Belief Networks to represent probabilistic interactions of accident precursors. The descriptive texts of individual causal factor nodes comprising the BBN of an accident are decomposed into its building blocks. By doing so an attribute-oriented classification and categorization system is introduced to represent individual causal factors. An attribute model is created based on this attribute-oriented classification and categorization system for each accident. The attribute models of separate accidents are merged to form a domain model representing the domain of interest of the CBR system. Finally, an initial seed based on 15 aviation accidents is created and indexed into the case base of the CBR system.