This paper presents an iterative method for deriving effective mesoscale potentials from atomistic simulations, applicable to polymer systems. The method uses differences in the potentials of mean force between distribution functions generated from a guessed potential and the true distribution functions to iteratively improve the effective potential. The algorithm is highly effective, converging quickly for any trial function. As a test case, the authors coarse-grained an all-atom model of poly(isoprene) (PI) using a 13:1 reduction of degrees of freedom, applying the method to both PI solutions and melts. The resulting force fields were compared, showing that a single force field cannot accurately represent different concentration regimes. The method is based on iterative Boltzmann inversion, which uses known correlation functions to derive effective potentials. It was tested on simple model liquids, demonstrating rapid convergence and the ability to reproduce target radial distribution functions (RDFs) accurately. The method was then applied to PI simulations, where it successfully reproduced structural details of the atomistic system with high accuracy. The resulting mesoscale model was validated against atomistic simulations, showing good agreement for both structural and dynamic properties. The method was also tested on different ranges of RDFs, showing that the iterative approach is robust and can handle various potential ranges. For the PI melt, the method was used to optimize the non-bonded interactions, leading to a potential that closely matched the target RDF. A pressure correction was also applied to improve the accuracy of the potential at long distances. The authors conclude that the iterative Boltzmann inversion method is a powerful tool for deriving mesoscale potentials from atomistic simulations, particularly for polymer systems. It is robust, converges quickly, and can accurately reproduce structural details of the system. The method is also applicable to other systems where intramolecular connectivity is important, such as low-molecular solvents and proteins. The results show that the method can successfully coarse-grain both melts and solutions, with the resulting potentials depending on the state of the polymer, here its environment. The method is recommended for use in polymer simulations, as it provides a reliable way to derive effective potentials from atomistic data.This paper presents an iterative method for deriving effective mesoscale potentials from atomistic simulations, applicable to polymer systems. The method uses differences in the potentials of mean force between distribution functions generated from a guessed potential and the true distribution functions to iteratively improve the effective potential. The algorithm is highly effective, converging quickly for any trial function. As a test case, the authors coarse-grained an all-atom model of poly(isoprene) (PI) using a 13:1 reduction of degrees of freedom, applying the method to both PI solutions and melts. The resulting force fields were compared, showing that a single force field cannot accurately represent different concentration regimes. The method is based on iterative Boltzmann inversion, which uses known correlation functions to derive effective potentials. It was tested on simple model liquids, demonstrating rapid convergence and the ability to reproduce target radial distribution functions (RDFs) accurately. The method was then applied to PI simulations, where it successfully reproduced structural details of the atomistic system with high accuracy. The resulting mesoscale model was validated against atomistic simulations, showing good agreement for both structural and dynamic properties. The method was also tested on different ranges of RDFs, showing that the iterative approach is robust and can handle various potential ranges. For the PI melt, the method was used to optimize the non-bonded interactions, leading to a potential that closely matched the target RDF. A pressure correction was also applied to improve the accuracy of the potential at long distances. The authors conclude that the iterative Boltzmann inversion method is a powerful tool for deriving mesoscale potentials from atomistic simulations, particularly for polymer systems. It is robust, converges quickly, and can accurately reproduce structural details of the system. The method is also applicable to other systems where intramolecular connectivity is important, such as low-molecular solvents and proteins. The results show that the method can successfully coarse-grain both melts and solutions, with the resulting potentials depending on the state of the polymer, here its environment. The method is recommended for use in polymer simulations, as it provides a reliable way to derive effective potentials from atomistic data.