Systematically coarse-grained (CG) force field models are state point dependent and therefore are usually not transferable to different chemistries and thermodynamic conditions. Fragment-based approaches for deriving CG polymer force fields increase the transferability of the models, but several challenges remain, in particular to describe conformational transitions of polymers driven by changes in the solvent environment or chemical composition of the chain. Herein, we describe a coarse-graining approach for deriving implicit-solvent CG polymer models, which can be used to study hydrophobic collapse transitions in aqueous solutions. On the basis of an earlier reported CG model for sodium poly(styrenesulfonate) (PSSNa) [Macromolecules 2012, 45, 2551−2561], we describe an approach to rescale nonpolar pair potentials of mean force based on a solvent-accessible-surface-area argument, thus allowing to reduce the fraction of charged monomers in the chain and to account for partial chain collapse. Based on a comparison with extensive detailedatomistic, explicit-solvent, simulations, it is demonstrated that the rescaling function for nonpolar interactions is chain length dependent. Extrapolation of the rescaling factor to long chains provides a new way to perform CG simulations of polymer solutions with moderately hydrophobic polyelectrolyte chains. Upon application to a 50%-charged PSSNa chain we observe pearl-necklace conformations in a cascade evolving manner with increasing chain length, in agreement with theoretical predictions for polyelectrolytes in poor solvent.