This research investigates a crime modelling technique known as crime scripting. Crime scripts are generated by crime analysts to improve the understanding of security incidents, and in particularly, the criminal modus operandi (i.e., how crimes occur) to help identify cost-effective crime prevention measures. This thesis makes four contributions in this area. First, a systematic review of the crime scripting literature that provides a comprehensive and up-to-date understanding of crime scripting practice, and identifies potential issues with current crime scripting methods. Second, a comparative analysis of crime scripts which reveals differences and similarities between the scripts generated by different analysts, and confirms the limitations of intuitive approaches to crime scripting. Third, an experimental study, which shows that the content of crime scripts is influenced by what scripters know about the future use of their scripts. And fourth, a novel crime scripting framework inspired from business process modelling and goal-based modelling techniques. This framework aims to help researchers and practitioners better understand the activities involved in the development of crime scripts, and guide them in the creation of scripts and facilitate the identification of suitable crime prevention measures.