AbstractTo achieve drug release from polymer prodrug nanoparticles, the drug‐polymer linker must be accessible for cleavage to release the drug, which can occur under certain physiological conditions (e.g., presence of specific enzymes). Supramolecular organization of polymer prodrug nanoparticles is crucial as it greatly affects the location of the linker, its surface exposure/solvation and thus its cleavage to release the drug. Since experimental access to these data is not straightforward, new methodologies are critically needed to access this information and to accelerate the development of more effective polymer prodrug nanoparticles, and replace the time‐consuming and resource‐intensive traditional trial‐and‐error strategy. In this context, we reported here the use of a coarse‐grained model to assist the design of polymer prodrug nanoparticles with enhanced cytotoxicity. By choosing the solvent accessible surface area as the critical parameter for predicting drug release and hence cytotoxicity of polymer prodrug nanoparticles, we developed an optimized polymer‐drug linker with enhanced hydrophilicity and solvation. Our hypothesis was then experimentally validated by the synthesis of the corresponding polymer prodrugs based on two different drugs (gemcitabine and paclitaxel), which demonstrated greater performances in terms of drug release and cytotoxicity on two cancer cell lines. Interestingly, our methodology can be easily applied to other polymer prodrug structures, which would contribute to the development of more efficient drug delivery systems via in silico screening.