Abstract: The tiered nursing care system constitutes a core standard for clinical nursing services, while the misalignment in nursing human resource allocation represents a critical bottleneck constraining healthcare quality. This review systematically examines recent advances in nursing workforce allocation models both domestically and internationally, analyzing the evolutionary patterns and key technological pathways underlying these models. Current allocation paradigms are shifting from experience-based staffing and fixed nurse-to-bed ratios toward data-driven approaches utilizing Patient Classification Systems (PCS). However, this transition faces challenges including a lack of standardized assessment tools, insufficient dynamic responsiveness, and inadequate information technology support. This paper focuses on synthesizing the current state of research concerning PCS-based allocation models. It summarizes the applicable scenarios, barriers to clinical translation, and limitations of three major mathematical modeling approaches—linear programming, queuing theory, and machine learning—and reviews methods for model validation and evaluation. Future research should prioritize developing patient classification tools aligned with China's national tiered nursing standards, advancing the standardization of nursing workload measurement, and exploring intelligent predictive models that integrate multisource data. Such efforts will provide essential theoretical and methodological support for establishing a precise, dynamic nursing human resource allocation system.