In innovative organizations, knowledge workers function as strategic assets and primary drivers of knowledge creation, transfer, and application, thereby playing a vital role in sustaining and enhancing competitive advantage. Consequently, their turnover intention can lead to adverse outcomes, including the loss of tacit knowledge, increased recruitment and training costs, and a reduced capacity for organizational innovation. This study aims to develop an innovative human resource management framework for predicting the turnover intention of knowledge workers in innovative organizations. To achieve this purpose, a qualitative meta-synthesis approach was employed to identify the dimensions and components influencing turnover intention. Through a systematic search of reputable scientific databases, 530 studies were initially screened, of which 31 eligible articles were selected and coded. The thematic synthesis of these texts resulted in the identification of 13 overarching dimensions and 36 subcomponents that collectively describe the multidimensional nature of turnover intention among knowledge workers. The credibility of the findings was enhanced using Lincoln and Guba’s four trustworthiness criteria. The resulting framework was subsequently applied in the parent study as the basis for feature engineering and the development of input variables, which were integrated into machine learning algorithms for predicting turnover intention. Accordingly, the findings of this research enrich the theoretical literature while providing a foundation for developing analytical and decision-support systems within data-driven human resource management.