Vacuum impregnation (VI) is a process of fluid replacement in porous media by reduction in atmospheric pressure and its subsequent reestablishment. The objective of this study was to evaluate the influence of the characteristic length and viscosity of impregnating fluids on vacuum impregnation dynamics. Refractory ceramic samples were used as a non-deformable porous media model, and a device continuously recorded the changes in net force (difference between weight force and buoyant force) on the sample through a load cell during the impregnation process. The relative values for the sample’s volumetric fraction due to spontaneous imbibition, for the drained fraction due to vacuum application, and for impregnation due to pressure reestablishment were estimated. The total volumetric fraction estimated during VI by the experimental device was compared with the values estimated by a balance and with those predicted by an equilibrium model. The experimental device was shown to be useful to determine impregnation kinetics, it was accurate and obtained values very close to the ones estimated by the balance and predicted by theoretical models. As expected, the kinetics was dependent on fluid viscosity and on the sample’s characteristic length. The kinetics data allowed the minimum time step at the VI process to be determined, enabling the optimization of the process applied to large media or in viscous fluid impregnation.