Abstract: 【Purpose/significance】Using knowledge elements as the basic unit, we perform semantic representation, semantic relation⁃ ship mining, and organizational aggregation on doctor-patient Q&A resources in online health communities to provide users with treatment-related health knowledge.【Method/process】Firstly, based on the existing knowledge element theory, a knowledge element semantic description model is constructed based on the structural characteristics of the Q&A texts; secondly, UIE model, external medical knowledge bases and pre-training model are used to identify, standardize and integrate entities in doctor-patient Q&A texts; then, utilize the relationships between entities and entity collections to mine dual semantic associations between knowledge elements; finally, based on the patient's actual condition, fine-grained and coarse-grained organization and aggregation of knowledge elements are carried out to meet the patient's knowledge needs.【Result/conclusion】This article starts from the perspective of knowledge ele⁃ ments, while ensuring the semantic integrity of knowledge units and considering the specific health context of patients, and conducts research in accordance with steps such as knowledge representation, entity extraction, semantic relationship mining, and organiza⁃ tional aggregation, so that the organized health knowledge can fully respond Condition needs of online health community users.【Inno⁃ vation/limitation】The theory of knowledge element is used to organize the knowledge of health Q&A information around the specific needs of users. However, in the construction process of the knowledge meta-semantic model, the doctor portrait information is not fully considered.