Polyethylene terephthalate (PET) waste is a common pollutant in the environment, mainly due to resistance of the plastic to bio-degradation. Nevertheless, hydrolytic enzymes have been identified with activity on this substrate, which are continually being engineered to increase activity. Some insoluble biological polymers are degraded by enzymes with a multi-domain architecture, comprising of a catalytic domain, and a substrate-binding domain, such as a carbohydrate-binding module (CBM). Enzymes that degrade PET have been shown to have a higher activity when fused with these CBMs, indicating a promising route for engineering better enzymes for plastic bioprocessing. However, no detailed study of the affinity and binding mechanism of these domains on PET has yet been made. Here, we perform an in depth analysis of a binding domain from CBM family 2 on PET, showing that the affinity of the protein for the plastic is highly dependent on temperature and crystallinity of the plastic. We also investigate the mechanism of the interaction, and show how affinity may be engineered in both directions. CBM affinity for other synthetic polymers is also demonstrated for the first time. Our results demonstrate that the substrate affinity of fusion enzymes with binding modules can be tuned to the desired level.