This paper presents a computational analysis of nicotine addiction using an adaptive network modelling paradigm. The model formalizes the contemporary theory of addiction as a failure of self-regulation, where epigenetic modifications of the CHRNA5 gene expression, progressively dismantles the prefrontal cortex’s (PFC) top-down control over behaviour. This creates a stable, self-sustaining addicted state characterized by impaired executive function and heightened cue-reactivity. Simulation experiments were performed for three scenarios: addiction onset following an environmental trigger, standard symptom-focused therapy, and a hypothetical epigenetic therapy. This work presents an integrative computational framework conceptualizing addiction as a chronic disease of impaired self-regulation, suggesting that effective long-term interventions must target these root biological mechanisms rather than surface-level symptoms.