Talent management has increased as a key factor in the success of organizations since the 1990s, along with the emergence of the concept of "war for talent". However, most of the research conducted has examined this concept at the level of human resource management and in a single-level manner; while numerous variables at the individual, organizational, and supra-organizational levels affect this process. The present study was conducted with the aim of designing a multi-level talent management model in the context of Iranian state-owned companies. In this regard, Tavanir Company was selected as one of the large and leading state-owned organizations. This study was conducted with a qualitative approach and a multi-method method; in a way that data-based theorizing (emergent approach) and multi-level analysis in a process manner were used. The participants were human resource managers and experts of Tavanir Company, who were selected using the theoretical sampling method. Data were collected through in-depth semi-structured interviews and analyzed using MAXQDA software version 2020. The main finding of the research is a multi-level model of talent management, which consists of five main components of talent management (as a central category), individual influencing variables, organizational influencing variables, supra-organizational influencing variables, and multi-level consequences. These components and the relationships between them were analyzed top-down, bottom-up, and horizontally, in the form of a dynamic process.