This article aims to propose a dynamic model of the propagation of political hate speech in the 2024 Aceh Regional Election using the SEIPR (Susceptible-Exposed-Infected-Paused-Recovered) epidemiological model approach modified for the socio-political context. S is a subpopulation that is not yet aware of the political hate speech (PHS) or is susceptible to receiving PHS; E is a subpopulation that is friends with PHS spreaders on social media and has tendency to support PHS; I is a subpopulation that supports and spreads PHS; P is a subpopulation that has spread PHS but stopped temporarily, but is possible to spread it again; and R is a subpopulation that no longer spreads PHS because i tis aware of the law and sanctions. This model analysis integrates one-at-a-time (OAT) sensitivity analysis and partial rank correlation coefficient (PRCC) to identify the parameters most influential on the peak number of spreaders (I) and the cumulative total of content, whose data is obtained from the Alliance of Independent Journalists (AJI) report regarding the hate speech on platform X (Twitter) and TikTok. The OAT results indicate that α and γᵢ have the greatest influence on the peak number of infections. The PRCC analysis results above confirm the OAT results, with the highest positive correlation values found in β and η, indicating that the higher the rate of social contact and transition to active spreaders, the greater the peak of the propagation of political hate speech. Scenario simulations indicate that overall, a comparison of the three scenarios shows that the intensity of the propagation of political hate speech is not only influenced by important political moments, but also highly sensitive to the type of narrative employed. The results of this study provide quantitative insights for the formulation of mitigation strategies, including digital literacy education, content monitoring, and community-based interventions to suppress the propagation of political hate speech during future political moments.