In this dissertation, we investigate how agent-based modeling (ABM) can contribute to understanding the complexity of radicalization. Radicalization is a complex and dynamic process, characterized by multiple pathways and outcomes. Although many contributing factors have been identified, there is a lack of integrated process models that explicitly explain how individual, social, and ideological mechanisms interact over time. ABM is well suited for developing such process models because it can explicitly represent interactions between individuals and their environment and explore their consequences through simulation. At the same time, there is a lack of guidelines for developing and validating such models. To address this, we introduce a set of design principles and a qualitative validation method for the development and evaluation of explanatory agent-based process models. These are applied throughout the remainder of the dissertation. We present the Individual, Group, Ideology (IGI) model of radicalization, which focuses on the interaction between individual, social, and ideological mechanisms. The central assumption is that these mechanisms can give rise to three interconnected processes: identity fusion, justification of extreme behavior, and ingroup superiority. Each of these processes constrains the range of perceived behavioral options, but none necessarily leads to radicalization on its own. When these processes occur simultaneously and reinforce one another, a situation emerges in which alternative courses of action gradually disappear. Under specific circumstances, this can lead to radicalization. The model is then operationalized through an iterative process. Simulations demonstrate the conditions under which radicalization may emerge. The results indicate that radicalization is a dynamic and relational process in which general social mechanisms can, under particular circumstances, lead to escalation. At the same time, the findings emphasize that radicalization is not an inevitable outcome but remains dependent on specific conditions and the sequence in which mechanisms become dominant. The findings demonstrate that the model connects explanatory theories and empirical observations by showing how radicalization emerges from mechanisms that reinforce one another through interaction. Finally, the model is evaluated by experts from the Dutch Police. The main benefits identified are knowledge integration, the ability to analyze radicalization from multiple perspectives, and the capacity to interpret movements and groups in terms of their dominant underlying mechanisms.