Abstract: Mikania micrantha is a pernicious invasive weed in agriculture and forestry. In this paper, the potential geographic distribution of Mikania micrantha in China was predicted by genetic algorithm for rule-set production（GARP) and maximum entropy (MaxEnt) models, and determined better model by receiver operating characteristic curve. Meanwhile, environmental variables were analyzed with jackknife method to judge the influence on Mikania micrantha. The results showed that the areas under ROC curves with GARP model and MaxEnt model were 0.910 and 0.971 respectively. It meant that MaxEnt model was better than GARP model in predicting the potential geographic distribution of Mikania micrantha. The results from jackknife method indicated that attitude and the variance of seasonality precipitation had little influence on the distribution of the invasive weed, while the range of annual temperature，annual precipitation, precipitation in the wettest month, precipitation in the wettest quarter and the variance of temperature range had greater influence on the distribution of the weed. According the prediction of MaxEnt, the potential geographic distribution areas of Mikania micrantha were in Hainan, Guangdong, Guangxi, Hong Kong, Macao, Yunnan, Fujian, Tibet and Guizhou. Southeast and southwest of Tibet, southwest of Guizhou and south of Fujian were areas which should be strengthened in monitoring and early warning work.