Understanding how competing narratives spread across social media requires models that account for user belief alignment and dynamic interactions. This paper introduces \(SEI_A I_D Z\) , a novel stance-aware epidemiological framework that categorizes users based on their narrative alignment: promoting and opposing. Unlike traditional information diffusion models that assume a single homogeneous narrative, \(SEI_A I_D Z\) simulates the simultaneous spread and contestation of divergent viewpoints within online environments. Using empirical data from TikTok during Taiwan’s 2024 presidential election, the model captures the behavioral transitions of users as they engage with politically charged content. Our analysis shows that stance-aware segmentation significantly improves prediction accuracy compared to classical models such as SEIZ. Key findings highlight the influence of transmission rate and stance-transition dynamics on narrative propagation. The \(SEI_A I_D Z\) model not only bridges computational and social theories but also offers actionable insights for platform governance, such as how content visibility and user behavior interact to shape information trajectories. This interdisciplinary approach provides a scalable and interpretable tool for researchers and policymakers seeking to analyze or intervene in online discourse. The framework lays the groundwork for future studies on belief-driven diffusion and the evolving dynamics of digital public opinion.