Ionic polymer-metal composites (IPMCs) consist of a perfluorinated ionomer membrane (usually Nafion® or Flemion®) plated on both faces with a noble metal such as gold or platinum and neutralized with the necessary amount of counterions that balance the electrical charge of anions that are covalently fixed to the backbone ionomer. IPMCs are electroactive materials with potential applications as soft actuators and sensors. Their electrical-chemical-mechanical response is dependent on the cations used, the nature and the amount of solvent uptake, the morphology of the electrodes, the composition of the backbone ionomer, the geometry and boundary conditions of the composite element, and the magnitude and spatial and temporal variations of the applied potential. Our most recent experimental results show that solvents can have profound effects on the nature of the IPMCs’ actuation. For example, we have discovered experimentally that Nafion-based IPMCs in Li+-form show very small back relaxation when hydrated, but extensive back relaxation with all other solvents that we have considered. On the other hand, the same membrane in the K+-form has extensive back relaxation when solvated with water, ethylene glycol, or glycerol, but none with 18-Crown-6. In the present paper, we seek to model the IPMCs’ actuation and compare results with the experimental data. The modeling rests on the observation that a sudden application of a step potential (dc) of several volts (1–3V) alters the distribution of cations within the ionomer, forcing cations out of the clusters near the anode and additional cations into the clusters near the cathode. The clusters within a thin boundary layer near the anode are thus depleted of their cations, while cations accumulate in the clusters near the cathode boundary layer. We first seek to determine the spatial and temporal variations of the cation distribution across the thickness of the IPMC for various cations and solvents, using an implicit finite difference numerical solution of the basic field equations, and compare the results with those of approximate analytical estimates. Based on this information, we then calculate the changes in the osmotic, electrostatic, and elastic forces that tend to expand or contract the clusters in the anode and cathode boundary layers. Finally, we calculate the amount of solvent out of or into the clusters that produces the bending motion of the cantilever. Comparing the model results with those of experimental measurement, we have arrived at remarkably good agreements. Indeed, our nanoscale-based model correctly predicts the unexpected influence of solvents on the actuation of IPMCs.