The development of efficient food packaging is essential to protect the product from moisture and gases and guarantee an adequate stability throughout its entire shelf life[1]. In the last decades, multiwall polymers have become widely used in food packaging due to their excellent mechanical and barrier properties. Unfortunately, the recycling of these materials is difficult and costly, leading to environmental problems[2]. An understanding of the structural and dynamic properties of polymers is crucial towards the development of new recycling technologies. However, the computational modeling of these materials poses a great challenge due to the fact that the dynamics span over several time scales[3]. As a result, a multiscale approach becomes mandatory to gain access to the slower degrees of freedom. In this regard, coarse grained (CG) models have drawn considerable attention in recent years[4,5]. CG models neglect the fastest degrees of freedom, flattening the rugged potential energy surface (PES) and thus, allowing higher time steps. In this work, we adopt a multiscale protocol to characterize several properties of polyethylene terephthalate (PET), which is commonly used in the food industry. First, atomistic molecular dynamics (MD) simulations are performed with a previously developed force field for amorphous PET systems with varying chain lengths. The structural properties are thoroughly examined as well as the glass transition temperature (Tg). Then, a CG force field is derived for several mapping schemes based on the iterative Boltzmann inversion (IBI)[6]. Finally, CG-MD simulations are performed to study the effect of the CG mapping on the structural properties.