Non-native invasive species establish in favourable habitats in alien regions. Such favourable habitats are largely determined by local climatic, soil and biogeographic factors. Modelling these factors can help managers to identify areas of possible risk of invasion. This paper uses logistic regression modelling to identify variables conducive to high invasion in a tropical mixed forest in a biodiversity hotspot region in India. Using presence-absence data of an invasive species Lantana camara and local habitat variables from increasing buffer distances around sampling locations along with broad scale climatic parameters, we identify the variables that support invasion and spread. Results indicated that the percentage of moist deciduous forest at a distance of 50 m around the plot was significantly related to the invasion of L. camara. The study demonstrates the facilitation by moist deciduous forests to the growth and spread of L. camara in this region, and highlights the importance of using data at multiplescales for modelling invasion.