In this study we apply an evidence-based approach to model population-size fluctuations and their corresponding impact on land use during the Roman and early-medieval periods in the Rhine–Meuse delta in the present-day Netherlands. Past-population numbers are reconstructed based on Roman and early-medieval settlement patterns. Corresponding impacts of these demographic fluctuations on potential land use are calculated by integrating the newly developed demographic overviews with archaeological and geoscientific data using a new land-use model termed ‘Past Land-Use Scanner’ (PLUS). The primary aims are to reconstruct first-millennium palaeodemographics and to explore the potential of simulation modelling for testing the feasibility of archaeological hypotheses regarding past land use. Results show that in the study area the first millennium AD was characterised by two periods during which major population growth occurred: the middle-Roman period (AD 70–270) and early-medieval period C (AD 725–950). A major demographic decline of 78–85% occurred during the late-Roman period (AD 270–450), after which first-millennium population numbers never again reached middle-Roman period levels. The modelling outcomes demonstrate that the impact of population fluctuations (growth vs decline) on the limits of the natural landscape during the first millennium in general was low. During these thousand years, the natural landscape almost without exception (only scenario D deviates) provided sufficient options for arable farming, meadows and pastures and was not a limiting factor for population growth. These results underline the added value of simulation modelling for testing the feasibility of archaeological hypotheses and analysing human–landscape interactions in the past.