In this paper, a mathematical model of regionalization based on graph theory to investigate the patterns induced by movements of livestock vehicles in cities under outbreaks of highly pathogenic avian influenza (HPAI) is proposed. We then compare the results of simulation from the regionalization model to actual HPAI outbreaks in 2016/2017 to evaluate the validity of the model. Specifically, we (1) configured a complex network structure with analytic tools and properties in graph theory to abstract the paths among farms and livestock facilities; (2) employed statistical methods to estimate the possibility of propagation between two clusters; (3) applied the developed method to an actual HPAI outbreak in Korea in 2016 and conducted a simulation to determine if the proposed modeling for regionalization is an effective prediction measure. The clustered regions proposed by the simulation correctly reflected the regional clustering of actual cases, while simultaneously contain the cities exposed to potential damage when separated. Based on these findings, we conclude that our proposed regionalization model is suitable for making policy judgments to establish a preemptive biosecurity system.