Abstract: 【Purpose/significance】To establish a new propagation model of network rumor, and provide useful reference for dealing with and managing network rumor.【Method/process】According to the propagation characteristics of network rumor, the CFDR propa‐gation model of network rumor was established by the principle of drug diffusion in pharmacokinetics, and the corresponding math‐ematical equations were given. Furthermore, considering the influence of authoritative media and time delay, the model is optimized,and the peak prediction and total number of rumor propagation are revised. The sensitivity of parameters is analyzed by using Matlab,Python and other tools. Finally, the "Chengdu 49 Middle School " event is taken as an instance to verify the validity of the model, and the relevant influencing factors are analyzed.【Result/conclusion】The results show that the CFDR model can be used to better fit the whole process of network rumor propagation, and all parameters of the model have an impact on the spread of network rumor. Among them, the positive intervention of authoritative media plays an important role in the spread of network rumor.【Innovation/limitation】The innovation is mainly reflected in: breaking through the previous inherent mode of studying network public opinion and rumor propagation based on infectious disease dynamics model, establishing CFDR network rumor propagation model by referring to drug dif‐fusion principle, and considering the intervention factors of authoritative media and the influence of time delay. The main limitation is as follows: only the actual data of one platform in one event is used in the instance verification, which can be further enriched in the fu‐
ture.