We compared estimates of crustal velocities within Great Britain based on continuous global positioning system (CGPS) measurements to predictions from a model of glacial isostatic adjustment (GIA). The observed and predicted values for vertical motion are highly correlated indicating that GIA is the dominant geodynamic process contributing to this field. In contrast, motion of the Eurasian plate dominates the horizontal motion component. A model of plate motion was adopted to remove this signal in order to estimate intraplate horizontal motion associated with GIA. However, a coherent pattern of horizontal motion was not evident in the resulting velocity field. We adopted a recently published model of the British–Irish ice sheet to predict vertical crustal motion for a large number of spherically symmetric Earth viscosity models. Our results show that the adopted ice model is capable of producing a high-quality fit to the observations. The CGPS-derived estimates of vertical motion provide a useful constraint on the average value of viscosity within the upper mantle. Values of model lithospheric thickness and lower mantle viscosity are less well resolved, however. A suite of predictions based on an alternative ice model indicates that the vertical motion data are relatively insensitive to uncertainties in the ice loading history and so the constraints on upper mantle viscosity are robust.