Shallow landslides can represent a destructive phenomena due to the absence of incipient movement evidence and to their high velocity, especially when turning into debris flows. The occurrence and the behavior of rainfall-triggered shallow landslides are often influenced by land use. In particular, the analysis of the changes in land use through time may represent a relevant indicator to evaluate different shallow landslide susceptibility scenarios. In this work, a semi-parametric nonlinear regression technique, namely the Generalized Additive Model (GAM), was applied for landslide susceptibility mapping. The model was applied to four different study sites which recently were affected by widespread rainfall-induced shallow landsliding: the Rio Frate and the Versa catchments (Oltrepo Pavese—southern Lombardy), the Vernazza catchment (Cinque Terre—Eastern Liguria) and the Pogliaschina catchment (Vara Valley—Eastern Liguria). Moreover, the elaboration of different land use scenarios were used to assess the influence of the land use change on the landslides occurrence.