Purpose/Significance To construct a satisfaction prediction model based on the dual perspectives of doctors and patients， and to provide references for the refined management of doctor-patient relationships in internet hospitals. Method/Process The pa. tients’emotions，communication behaviors，the use of sensitive words，and doctors’response patterns and other features are dynami. cally captured. Combined with affective computing and historical interaction data，communication satisfaction prediction and dynamic assessment of potential risks are achieved. Result/Conclusion The proposed model can effectively identify potential high-risk communi. cation scenarios，providing an implementable technical solution for the pre-identification and proactive intervention of dissatisfactionamong patients in internet hospitals.