This study aimed to develop a dynamic model of neonatal mortality in Kerman and Bam cities, Iran, to identify the complex interactions affecting neonatal health outcomes and to propose effective strategies for healthcare policy and management. Utilizing a systems dynamics approach, we first constructed Causal Loop Diagrams (CLDs) to depict the qualitative interactions among factors influencing neonatal mortality, which were subsequently transformed into Stock and Flow Diagrams for quantitative analysis. Data were collected over a 60-month period (2017 to 2021) from the Iranian Maternal and Neonatal (IMaN) database and the Integrated Health System (SIB), and were supplemented with expert interviews and hospital informatics to enhance data robustness. The developed model demonstrated high validation accuracy, achieving approximately 94% based on Mean Absolute Percentage Error (MAPE) when compared with historical data. Key determinants of neonatal mortality were categorized into health factors (e.g., preterm birth, eclampsia), socio-demographic factors (e.g., maternal education, substance abuse), and healthcare system factors (e.g., NICU capacity, specialist staff). Simulation scenarios indicated that targeted interventions in these areas could significantly reduce neonatal mortality rates. Consequently, the presented dynamic model offers a robust framework for simulating healthcare scenarios and making data-driven decisions in resource allocation, aiding healthcare managers and policymakers in optimizing interventions such as increasing NICU capacity and improving maternal care programs. Future research should focus on budget optimization, developing adaptive algorithms, and creating digital twins of the healthcare system to facilitate broader macro-level analyses.