Monoethylene glycol (MEG) injection is an approved technology to displace the hydrates equilibrium loci in ultra-deep subsea gas production pipelines. Further, rich MEG has to be processed in MEG Recovery Units (MRU) in order to be recovered as lean MEG to be pumped back. As distillation-based technologies, MRUs are very intensive in terms of heat consumption. In offshore platforms, where space and resources are limited, besides determining energy requirements, it is also of importance to perform a thermodynamic analysis to assess degradation of energy quality, which is done by Exergy Analysis (ExA). The exergy flow of a stream is the maximum amount of power obtainable when the stream is brought into equilibrium with a reference environmental reservoir (RER). In this regard, exergy is a property that depends both on the state of the stream and the definition of the RER. This work investigates how the choice of MEG state in the RER impacts the Exergy Analysis of MRUs. Specifically, two choices of MEG states in the RER are compared by performing ExA for three technologies of MRUs: Traditional, Full-Stream and Slip-Stream. Both approaches led to consistent results with the Traditional MRU having the highest exergy efficiency (if MEG is not supposed to be degraded), while the Full-Stream MRU presented the lowest exergy efficiency, and the Slip-Stream MRU occupied an intermediate position. However, Conception #1 allowed exergy flows to be too high, masking the exergy losses. As a consequence, all MRUs have very similar exergy efficiencies. On the other hand, Conception #2 was able to better discriminate the exergy losses and could differentiate MRUs better. Only with the RER definition via Conception #2 it was possible to obtain more realistic results of exergy efficiencies.