The paper presents an iterative method for potential inversion from distribution functions, originally developed for simple liquid systems, and extends it to polymer systems. The method uses the differences in the potentials of mean force between distribution functions generated from a guessed potential and the true (simulated) distribution functions to iteratively improve the effective potential. The optimization algorithm is highly efficient, converging quickly for each trial function. The method is tested on a coarse-grained model of poly (isoprene) (PI) using a 13:1 reduction of degrees of freedom, both in solutions and melts. The resulting coarse-grained (CG) force fields are shown to be effective in reproducing the structural details of the atomistic system with high accuracy. The study also demonstrates that a single force field cannot be used for different concentration regimes, highlighting the importance of considering the chemical nature of the polymer in the coarse-graining process. The method is validated through convergence tests on model liquids and applied to the coarse-graining of PI, showing that the CG potentials depend on the state of the polymer, such as its environment.The paper presents an iterative method for potential inversion from distribution functions, originally developed for simple liquid systems, and extends it to polymer systems. The method uses the differences in the potentials of mean force between distribution functions generated from a guessed potential and the true (simulated) distribution functions to iteratively improve the effective potential. The optimization algorithm is highly efficient, converging quickly for each trial function. The method is tested on a coarse-grained model of poly (isoprene) (PI) using a 13:1 reduction of degrees of freedom, both in solutions and melts. The resulting coarse-grained (CG) force fields are shown to be effective in reproducing the structural details of the atomistic system with high accuracy. The study also demonstrates that a single force field cannot be used for different concentration regimes, highlighting the importance of considering the chemical nature of the polymer in the coarse-graining process. The method is validated through convergence tests on model liquids and applied to the coarse-graining of PI, showing that the CG potentials depend on the state of the polymer, such as its environment.