Abstract:Against the backdrop of digital and intelligent transformation, the integration and sharing of nursing data resources have become crucial pathways to improving the quality of evidence-based nursing.However, nursing data are often fragmented across various medical institutions and care settings.This exacerbates the prominent issue of data silos, restricting the in-depth advancement of multicenter nursing research.Federated learning enables multi-party collaborative computation and joint analysis without requiring raw data to leave local sites, offering a novel approach to the secure utilization of nursing data.Focusing on the data silo dilemma in multicenter nursing research, this paper discusses the applicability, technical principles, and primary models of federated learning.It reviews its applications in the development of clinical prediction models, chronic disease management, transitional care, and predictive nursing quality evaluation.Furthermore, the paper analyzes the current challenges faced in data quality, technical deployment, and ethical governance, aiming to provide a comprehensive reference for the future application of federated learning in multicenter nursing research.