Recent avian-origin H3N8 influenza A virus (IAV) that have infected humans pose a potential public health concern. Alterations in the viral surface glycoprotein, hemagglutinin (HA), are typically required for IAVs to cross the species barrier for adaptation to a new host, but whether H3N8 has adapted to infect humans remains elusive. The observation of a degenerative codon in position 228 of HA in human H3N8 A/Henan/4-10/2022 protein sequence, which could be residue G or S, suggests a dynamic viral adaptation for human infection. Previously, we found this human-isolated virus has shown the ability to transmit between ferrets via respiratory droplets, with the HA-G228S substitution mutation emerging as a critical determinant for the airborne transmission of the virus in ferrets. Here, we investigated the receptor-binding properties of these two H3N8 HAs. Our results showed H3N8 HAs have dual receptor-binding properties with a preference for avian receptor binding, and G228S slightly increased binding to human receptors. Cryo-electron microscopy structures of the two H3N8 HAs with avian and human receptor analogs revealed the basis for dual receptor binding. Mutagenesis studies reveal that the Q226L mutation shifts H3N8 HA´s receptor preference from avian to human, while the G228S substitution enhances binding to both receptor types. H3N8 exhibits distinct antigenic sites compared to H3N2, prompting concerns regarding vaccine efficacy. These findings suggest that the current H3N8 human isolates are yet to adapt for efficient human-to-human transmission and further continuous surveillance should be implemented.IMPORTANCEInfluenza virus transmission remains a public health concern currently. H3N8 subtype influenza A viruses infect humans and their HAs acquire the ability to bind to both human and avian receptors, posing a threat to human health. We have solved and analyzed the structural basis of dual receptor binding of recently human-infecting H3N8 HA, and we demonstrate that the G228S enhances human receptor binding and adaptation. We also found that HN/4-10 H3N8 HA has distinct antigenic sites, which challenges vaccine efficacy. Taken together, our work is critical to the prevention and control of human H3 influenza virus infection.