Abstract: The research on micro blog communication prediction was a hot spot. Micro blog forwarding was one of the key mechanisms. On the basis of traditional infectious diseases model, an improved SIRE model was proposed through integrating into the foreign user impact factor. Comparing with the SISe model, SIRE model could be used to fit well and forecast the micro blog forwarding rule. Comparing with Harmony Algorithm, the Particle Swarm Algorithm and Genetic Algorithm, Harmony Algorithm was used for parameters optimization of micro blog forwarding prediction model, its performance was the best, it could characterize the tread of micro blog communication.