In this paper, several factors are analyzed in the evolution of Internet public opinion in emergent events. The analysis was conducted with an illustration about the information spreading in “H7N9 Event” with the help of NetLogo. After agent-based modeling and simulation, some interesting results have been found: (1) compared to the Internet public opinions without government intervention, the situation with government interventions had a shorter time to reach stable state; (2) compared to the Internet public opinions with lower credibility of the government, the situation with higher credibility of the government had a shorter time to reach stable state; (3) the sooner the government disclosed information on the Internet, the sooner the public opinions reached stable state; (4) the more opinion leaders participated in the Internet public opinions, the earlier the opinions reached stable state.