Attempts to model insurgency have suffered from several obstacles. Qualitative research may be vague and conflicting, while quantitative research is limited due to the difficulties of collecting sufficient data in war and inferring complex relationships. We propose an innovative combination of Fuzzy Cognitive Maps and Cellular Automata to capture this complexity. Our approach is computational, thus it can be used to develop a simulation platform in which military and political analysts can test scenarios. We take a step-by-step approach to illustrate the potential of our approach in a population-centric war, similar to the on-going campaign in Afghanistan. While the project still requires validation and improvement of the knowledge base by domain experts as well as construction of accurate simulation scenarios, this example fully specifies the general problem definition and the technical structure of the model.