The factors underlying the invasiveness of non-native plants species and the invasibility of habitats in non-native ranges have been investigated in a number of studies. However, there are few examples where invasion potential across a range of key, vulnerable habitats within the non-native range has been compared for a single problematic species. Understanding the invasion potential of a species in different habitat types is crucial in prioritising management and control efforts, and in the protection of vulnerable habitats through monitoring. Here, using the invasive shrub Rhododendron ponticum as a case study, we integrate information on both the demographics and spatial dynamics within an individual-based, spatially-explicit model to investigate invasion potential in different habitats. Firstly, empirical demographic data were used to establish relationships between demographic traits, such as height and fecundity, and habitat variables. The outputs from models fitted using a Generalised Linear Model approach were then incorporated into an individual-based simulation model of plant spread to investigate the invasion potential in different habitats using a factorial design of treatments. Plant height, and thus seed release height, was found to be the main driver of invasion speed through an increase in dispersal potential, which resulted in the highest invasion speeds predicted for evergreen woodlands, and relatively low speeds for open habitats. Conversely, invasion density was driven by plant fecundity and seedling survival and not dispersal potential, which resulted in the highest invasion densities predicted for open habitats, and relatively low densities for evergreen habitats. Deciduous woodland had features that resulted in intermediate invasion potential, both in terms of speed and density and may, therefore be the habitat that is most vulnerable to rapid and dense invasion.