Abstract Aim A species’ dispersal characteristics will play a key role in determining its likely fate during a period of environmental change. However, these characteristics are not constant within a species – instead, there is often both considerable interpopulation and interindividual variability. Also changes in selection pressures can result in the evolution of dispersal characteristics, with knock-on consequences for a species’ population dynamics. Our aim here is to make our theoretical understanding of dispersal evolution more conservation-relevant by moving beyond the rather abstract, phenomenological models that have dominated the literature towards a more mechanism-based approach. Methods We introduce a continuous-space, individual-based model for wind-dispersed plants where release height is determined by an individual’s ‘genotype’. A mechanistic wind dispersal model is used to simulate seed dispersal. Selection acts on variation in release height that is generated through mutation. Results We confirm that, when habitat is fragmented, both evolutionary rescue and evolutionary suicide remain possible outcomes when a mechanistic dispersal model is used. We also demonstrate the potential for what we term evolutionary entrapment. A population that under some conditions can evolve to be sufficiently dispersive that it expands rapidly across a fragmented landscape can, under different conditions, become trapped by a combination of limited dispersal and a large gap between patches. Conclusions While developing evolutionary models to be used as conservation tools is undoubtedly a challenge, we believe that, with a concerted collaborative effort linking the knowledge and methods of ecologists, evolutionary biologists and geneticists, it is an achievable aim.