Accident causation models mainly answer the following two questions: (i) why does an accident occur, and (ii) how does it occur? These models are the most important theoretical basis for safety science, and provide an important method for accident analysis and prevention. To understand accident causation models systematically and comprehensively, this work clarifies the development history of these models over the past 100 years. The work conducted in this study is summarised as follows: (i) The role and origin of accident causation models are introduced. (ii) A new method for classifying accident causation models is proposed. The method divides the accident causal models into linear and nonlinear accident causation models, and the latter are further divided into human-based, statistics-based, energy-based, and system-based accident models. (iii) A review of 29 representative accident causation models proposed in the past 100 years is conducted. The theoretical basis, application flow, and application status of these models are highlighted. (iv) A detailed introduction to the 24Model, an accident causation model with theoretical innovation and more modern safety management, is presented. (v) A comparative analysis of various accident causation models and their development trends are discussed. (vi) This safety also summarises the application status of the accident causal model and prospects for future applications. The research findings of this study are as follows: (i) The newly proposed classification method of accident causation models clarifies the classification of accident causes. (ii) Each type of accident causation model has its own characteristics and application scope. In an accident analysis, an accident model that meets its industry characteristics should be selected. (iii) Organisational factors will be replaced by more modern safety management systems, and people will pay more attention to the role of safety culture in accident prevention. Accident causation models will develop in a linear and systematic way. (iv) The current accident causation models consist mainly of qualitative analysis and quantitative analysis, and will develop in the direction of dynamic analysis, accident prediction, and intelligent comprehensive analysis in the future. (C) 2019 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.