Much progress has been made in understanding the dynamics of spontaneous epileptic seizures [1], while an understanding of why and how the disease develops (epileptogenesis) remains elusive. Here, we present a first-of-its-kind model of epileptogenesis describing the interactions between neuroinflammation, blood-brain barrier disruption, neuronal death, circuit remodeling, and seizures. The model is formulated as a system of nonlinear differential equations describing processes acting on realistic timescales ranging from seconds to weeks. Mathematically, it is characterized by two stable fixed points corresponding to healthy and epileptic conditions divided by a separatrix. The model allows for simulation of epileptogenesis caused by various types of injuries, and captures characteristic injury-specific courses of pathology development. This was tested with data from 3 distinct animal models, in which epilepsy is triggered by neural infection, prolonged seizure (status epilepticus), or blood-brain barrier leakage. In addition, our model captures such features of epileptogenesis as the so-called latent period (time interval between injury onset and first seizure), the emergence of long timescales (up to decades) of pathology development, dose-dependence of epileptogenesis risk and severity on injury intensity, and variability of epileptogenesis outcomes in subjects exposed to identical injury. Furthermore, the model highlights the multicausal nature of epileptogenesis, showing that neuronal death is not necessary for epileptogenesis, while it is alone sufficient to cause disease development. Finally, our model predicts efficient injury-specific therapeutic strategies in the form of specific intervention time windows. In conclusion, our model provides new insights into the multi-causal nature of disease development and generates testable predictions for future experiments and therapeutic interventions.