Worms have posed a serious threat to Internet. Meanwhile, worms have posed a more serious threat to P2P networks based on Internet. The key properties of P2P networks, which bring facilities to users, result in vulnerabilities to attacks from P2P worms. Considering that models of the other two P2P worms of three classes of P2P worms，namely passive worms and active worms, have been proposed and reactive worms have seriously threatened P2P networks, the models of propagation and immunization of P2P reactive worms are proposed. Furthermore, the condition of worm free in the stead state is deduced from the model of propagation of reactive worms. In order to validate the epidemic model proposed and the condition of worms free in the steady state, large scale simulation experiments are carried out with the software Matlab. All the simulations validate the model and the necessary conditions. In addition, all the simulations show that it is the prevalent index that is in charge of whether worms are prevalent and the degree of worm spread if it will be prevalent. Obviously, the prevalent index is very helpful to find worm-throttling strategies. Analysis of the P2P-related parameters, which determine the value of the prevalent index, shows that decreasing the download rate is the most effective method of throttling worm spread before the corresponding patches are released.