Modeling the potential distribution of invasive plant species to implement effective prevention strategies is one of the major issues confronting rapidly urbanizing regions. This study focused on Mikania micrantha, the most problematic weed in the study site (Baoan District in Shenzhen, China). Our goal was to determine the key impact factors associated with the weed's presence/absence information through the comparison analysis between the invasive and non-invasive sites as well as the construction of Autologistic regression model. Data analysis was based on the land-use classification map derived from IRS satellite imagery of 2007 and contemporary survey map of Mikania, while topographic data were obtained using DEM in a geographic information system (GIS). The final conclusions are drawn from the research as follows:(1) At the regional scale, most topography and land use characteristics were significantly correlated with Mikania presence, whereas features of the local vegetation community revealed little influence; (2) Autologistic regression modeling demonstrated that the weed distribution was highly correlated with surrounding orchard density and water density, and this model showed a good performance of fitness, therefore, it could be used as a valuable tool for reconstructing the invasion process and assisting decision makers to target the locations at highest risk in the near future.