Demographic comparisons between wild and restored populations of at-risk plant species can reveal key management strategies for effective conservation, but few such studies exist. This paper evaluates the potential restoration success of Alyxia stellata, a Hawaiian vine. Stage-structured matrix projection models that compared long-term and transient dynamics of wild versus restored A. stellata populations, and restored populations under different levels of canopy cover, were built from demographic data collected over a four year period. Stochastic models of wild populations projected stable or slightly declining long-term growth rates depending on frequency of dry years. Projected long-term population growth rates of restored populations were significantly higher in closed than open canopy conditions, but indicated population decline under both conditions. Life table response experiments illustrated that lower survival rates, especially of small adults and juveniles, contributed to diminished population growth rates in restored populations. Transient analyses for restored populations projected short-term decline occurring even faster than predicted by asymptotic dynamics. Restored populations will not be viable over the long term under conditions commonly found in restoration projects and interventions will likely be necessary. This study illustrates how the combination of long-term population modelling and transient analyses can be effective in providing relevant information for plant demographers and restoration practitioners to promote self-sustaining native populations, including under future climates.