In nature, subducting slabs and overriding plate segments bordering subduction zones are generally embedded within larger plates. Such large plates can impose far‐field boundary conditions that influence the style of subduction and overriding plate deformation. Here we present dynamic laboratory models of progressive subduction in three‐dimensional space, in which the far‐field boundary conditions at the trailing edges of the subducting plate (SP) and overriding plate (OP) are varied. Four configurations are presented: Free (both plates free), SP‐Fixed, OP‐Fixed, and SP‐OP‐Fixed. We investigate their impact on the kinematics and dynamics of subduction, particularly focusing on overriding plate deformation. The results indicate that the variation in far‐field boundary conditions has an influence on the slab geometry, subduction partitioning, and trench migration partitioning. Our models also indicate that in natural (narrow) subduction zones, assuming a homogeneous overriding plate, the formation of back‐arc basins (e.g., Tyrrhenian Sea, Aegean Sea, and Scotia Sea) is generally expected to occur at a comparable location (250–700 km from the trench), irrespective of the boundary condition. In addition, our models indicate that the style of fore‐arc deformation (shortening or extension) is influenced by the mobility of the overriding plate through controlling the force normal to the subduction zone interface (trench suction). Our geodynamic model that uses the SP‐OP‐Fixed setup is comparable to the Calabria subduction zone with respect to subduction kinematics, slab geometry, trench curvature, and accretionary configuration. Furthermore, the model can explain back‐arc and fore‐arc extension at the Calabria subduction zone since the latest middle Miocene as a consequence of subduction of the narrow Calabrian slab and the immobility of the subducting African plate and overriding Eurasian plate. This setting induced strong trench suction, driving fore‐arc extension, and forced subduction to be accommodated almost entirely by slab rollback (not trenchward subducting plate motion), while trench retreat was accommodated almost entirely by back‐arc and fore‐arc extension (not trenchward overriding plate motion), comparable to our SP‐OP‐Fixed model.