Habitat loss, fragmentation and degradation are important factors in the decline of biodiversity worldwide. It is important to be able to evaluate the success of policies at different levels, including, increasingly, the global level. Whilst attention has been given to the development of predictive models that focus on individual species within biogeographic regions or smaller areas, however, to assess the impact of land-use change and policy measures on biodiversity at global level, there is an urgent need for generic tools (models, algorithms, databases). In this paper we test the potential of a generic tool, as part of the GLOBIO model, for assessing the impact of habitat loss and fragmentation. It combines existing data for the minimum viable populations of terrestrial bird and mammal species with knowledge of individual area requirements to derive estimates of their minimum area requirements (MAR). This approach focuses on comparing the minimum area requirements (MAR) to the natural habitat areas, assuming that below a certain threshold populations are no longer viable and the species assembly will eventually be reduced. The relationship between nature area and percentage of species meeting Minimum Area Requirements appears to be log-linear between 10 km2 and 10 000 km2 for conservation priority species and has the form Y=-15.45 + 28.61* LOG(AREA). Our results suggest that many existing parks and reserves might be too small for the long-term viability of species that they are meant to preserve. Applying this relationship to a global land cover dataset reveals that substantial proportions of mammal and bird species occur in areas that fail to cover sufficient space to support long term viable populations. This applies even at current state, especially for those areas of the globe where rapid urbanisation and agricultural expansion have taken place and are anticipated to proceed.