Rumor is an important form of social interaction. Therefore, spreading harmful rumors can have a negative impact on the health of the society. People's communication in the society plays an important role in spreading rumors, and whether or not it is spread depends on the person's level of trust in the rumor. Thus, one of the most important factors in a person's trust (or distrust) of a rumor is the number of neighbors who believe the rumor and spread it (and vice versa, the number of neighbors who do not believe the rumor and react to it). In this paper, we present this case in the form of linguistic variables and the use of fuzzy logic. In this paper, we propose an epidemic model of rumor dissemination in online social networks in which in addition to existing (susceptible–infected–recovered) modes, the rumor delay mechanism (exposed) is also added a counter attack mechanism (counterattack). The proposed model is presented as: susceptible–exposed–infected–counterattack–vaccinated–recovered–susceptible considering that the network and exposed node are constructed fuzzy. Using numerical simulations, we verify the performance of model in a SFN and a real network topology (Facebook). The simulation results of the proposed model show that compared to the SIRS and SEIRS models, the emission rate is lower, and the pollution is eliminated earlier.