This paper first considers population aging as a breakthrough, analyzes the demand for long-term care due to population aging, and classifies the models and characteristics of long-term care. Secondly, the combinatorial optimization algorithm and its classical cases are analyzed, and a care resource allocation system is built based on the solution set performance metrics and multi-objective combinatorial optimization algorithm for an aging population society. In city A, a research experiment examines the elderly population and the allocation of nursing care resources. The results show that the Gini coefficient of the population distribution of nursing institutions is 0.027 and the Gini coefficient of the population distribution of the number of institutional beds is 0.296, both of which are less than 0.3, thus indicating that the overall structure of the current social nursing resource allocation is relatively fair.