computer simulation is an effi cient tool. The main aim of such simulations is to probe the system at length and timescale which allows the study of the desired properties, while utilizing minimal computer resource. In this context, generic coarse-grained (CG) models [ 4 ] are quite useful as they aid in the study of specifi c properties with minimal reference to the molecular level picture. Generic models such as bead-spring, freely jointed, and worm-like chain models have been widely used to study polymer melts and solutions, and are capable of leading to general conclusions on their properties. On the other hand, parameterized model Martini [ 5 ] is extensively used in biomolecular simulations, specifi cally in the study of structure and dynamics of lipid membranes. While the former set of models apply to any polymer in general, Martini coarse-grains a molecular system (a specifi c protein, lipid, etc.). ChemicalFor mesoscale structural studies of polymers, obtaining maximum level of coarse-graining that maintains the chemical specifi city is highly desirable. Here we present a systematic coarse-graining study of sulfonated poly(ether ether ketone), sPEEK, and show that a 71:3 coarse-grained (CG) mapping is the maximum possible map within a CG bead-spring model. We perform single chain atomistic simulation on the system to collect various structural distributions, against which the CG potentials are optimized using iterative Boltzmann inversion technique. The potentials thus extracted are shown to reproduce the target distributions for larger single chains as well as for multiple chains. The structure at the atomistic level is shown to be preserved when we back-map the CG system to re-introduce the atomistic details. By using the same CG mapping for another repeat unit sequence of sPEEK, we show that the nature of the effective interaction at the CG level depends strongly on the polymer sequence and cannot be assumed based on the nature of the corresponding atomistic unit. These CG potentials will be the key to future mesoscopic simulations to study the structure of sPEEK based polymer electrolyte membranes.