Purpose/Significance The paper discusses the construction of the current elderly health policy system by using policy text mining and quantitative evaluation, and provides references for future policy adjustment and optimization. Method/Process Policies on elderly health issued by the relative departments from 2019 to 2023 are analyzed. The ROST text data mining tool is used to extract high-frequency words and build semantic networks. Gephi is utilized to visualize social network relationships for text analysis. Quantitative analysis of these policies is conducted using the PMC index model. Results/Conclusion The high-frequency words such as “national”,“society”,“knowledge”,“resources”, and “guarantee” are identified as key areas in elderly health policies, which are interconnected within the social network. The overall trend of the elderly health policy formulation is positive. It is suggested to improve the accuracy and adaptability of the policy, clarify the time node and core elements of the policy, and enhance the strategic analysis of the function and content of the policy.