Seasonal influenza activity significantly decreased in China during the coronavirus disease 2019 (COVID-19) pandemic, yet the H3N2 virus led to three epidemic waves. Understanding the characteristics of H3N2 epidemic viruses is essential for recognizing influenza during COVID-19 and for updating vaccines. In this study, we analyzed 579 respiratory samples from patients exhibiting influenza-like symptoms, collected in 2019–2022, leading to the successful sequencing of 36 complete H3N2 genomes. Genomic analysis indicated that the epidemic strains from these periods belonged to different hemagglutinin (HA) clades and exhibited phylogenetic divergence from the concurrently used vaccine strains. Significant antigenic differences were identified through cross-hemagglutination inhibition (HI) and cross-microneutralization (MN) assays. Furthermore, pathogenicity studies showed that representative strains replicated in Madin-Darby canine kidney‌ (MDCK) cells, with varying abilities, and all replicated more effectively at 37 °C compared to 33 °C. These strains also replicated well in the respiratory tracts of mice and guinea pigs. The findings indicate a mismatch between circulating H3N2 viruses and recommended vaccine strains, highlighting the need for improved international cooperation and epidemiological surveillance of influenza viruses post-COVID-19. Optimizing effective vaccine strain update strategy and developing a universal influenza vaccine are crucial for future preparedness.