Since early 2020, COVID-19 has been a major public security crisis that has had an enormous impact on the world. With the spread of the epidemic, rumors occur, some of which have even caused public panic. They have greatly affected the government’s efforts of epidemic prevention and thus urgently need to be evaluated. This study aimed to examine how to make flexible use of different policy tools to govern rumors based on their different characteristics. From the perspective of behavioral public policy, this study observes the effectiveness of various behavioral policy tools in rumor governance, hoping to explore the optimal solution of rumor governance from the perspective of micro public psychology. The survey experiment shows that individual behavior-related rumors (hereafter referred to as IBRs) are easier to be governed than epidemic progress-related rumors (hereafter referred to as EPRs) are, and that quick response is more effective than non-quick response. Through interaction analysis, it is known that in the governance of IBRs, nudge is more effective in rapid response, while in the context of non-quick response, boost outperforms nudge in rumor governance. A similar phenomenon can be seen in the scenario of EPR governance, despite a tinier difference in effectiveness compared with that of IBRs. The study enlightens us that rumor refutation requires not only people’s disbelief in and restraint on rumors, but also the implementation of science-based targeted policies. Based on the conclusion, this study puts forward suggestions on implementing targeted policies of rumor governance.