We are focusing on the effect of previous occupation on settlement pattern. The basic concept used is the ‘memory of landscape’, considering that occupation of archaeological settlements also reflects human investment in the surrounding area. As soon as rural communities settled somewhere, they reshaped the landscape by delimiting parcels, making clearances, and improving the quality of soils. We can therefore assume that the duration of rural settlement occupation is an index of long-term land use intensity. In this paper we will outline a spatial model to compute a map of land use intensity per time slice of one century. The aim is to include this map as a variable in predictive modelling, in order to estimate the weight of social investment in the landscape and its effect on subsequent settlement location choice. The issue is to estimate whether humans are more or less opportunistic when finding a new place to settle, and to understand why, also considering other variables.