Influenza A viruses (IAVs) represent a persistent threat to global public health because of their capacity to cross species barriers and rapidly adapt to new hosts. Post-translational modifications (PTMs) are known to regulate viral protein function, yet the evolutionary dynamics of PTM sites across influenza strains remain largely uncharacterized. In this study, we employed MusiteDeep to computationally predict potential PTM sites across influenza A virus proteins from H1N1, H5N1, and H7N9. We modeled predicted PTM states as discrete evolutionary traits using Bayesian phylogenetic methods to examine patterns of evolutionary rate variation at these computationally identified sites. Our analysis identified 34 positions at predicted PTM sites showing either significantly elevated (11 sites) or reduced (23 sites) evolutionary rates relative to other PTM-associated positions within each protein. Fast-evolving sites were enriched in polymerase proteins and surface glycoproteins, where slowly evolving sites were more broadly distributed, with notable concentrations in PB1 polymerase and NS1 protein. A subset of these sites showed potential host-associated rate differences, suggesting that selective pressures on PTM sites may differ across host lineages.