Land evaluation has a leading role in the sustainable management of natural resources. By integrating information from different disciplinary fields and at different spatial scales, concerning soils, climate, vegetation, geomorphology, economic and social drivers, it assesses alternative land uses with consideration of socio-economic and environmental objectives. Over time, the increasing complexity of such an analysis has shown the limits of traditional approaches proposed by Food and Agricultural Organisation and other international institutions in the second half of the 20th century, to deal with relatively new phenomena, such as those related to global change. Among the recent methodological proposals, agro-ecological zoning (AEZ) has gained increasing attention of scholars. Global change affects all the variables to be considered for land evaluation and in particular those affecting land productivity and economic consequences, but only rarely they have been jointly considered. Moreover, the possibility of simulating agro-ecosystems over long-time periods in many parts of the world is limited by the great efforts required for data acquisition and the many sources of errors. Our study aims to demonstrate the opportunity to explore the use of virtual territories as analogues of controlled field experiments, in order to carry out scenario analysis of agro-ecosystems, which may exist or not at the moment. Virtual territories exhibit morphological, ecological and land cover features statistically similar to selected existing territories. Provided this basic prerequisite is met, a virtual territory is built as a coherent set of geographic information system layers and databases on which the effects of global change phenomena (e.g., climate and land use changes) can be simulated under different scenarios. The advantages are: the possibility of having full control of the experimental conditions; the possibility of setting up factorial experiments with combinations of different typologies or levels of climate, physical conditions, socio-economic development, etc.; the efficiency and the flexibility of the tools adopted to easily generate realistic landscape and their variants. The approach is demonstrated through the development of erosion analysis under climate change scenarios.