Influenza viruses cause millions of infections worldwide each year. Influenza viruses constantly acquire mutations in their surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), through a process called antigenic drift. HA is the main target of anti-influenza neutralizing antibodies (Abs). Due to antigenic drift, the seasonal influenza vaccines must be updated annually to include the most antigenically relevant strains. Each year, the World Health Organization collects thousands of clinical influenza isolates, propagates them in cell culture, and performs both sequencing and serological analyses to assess the antigenic characteristics of circulating viral strains. In this dissertation, we investigate multiple factors associated with surveillance and vaccine strain selection that could be improved to produce more reliable and effective seasonal influenza vaccines. We first demonstrate that recent H3N2 subtype viral isolates rapidly acquire mutations in both HA and NA when propagated in cell culture, resulting in increased receptor binding avidity or NA-dependent receptor binding, respectively. These mutations impact antigenic analyses that are routinely used for viral surveillance. We then explore how a single mutation in HA antigenic site B contributed to the antigenic drift and subsequent vaccine mismatch of newly emerged H3N2 viruses during the 2014-2015 influenza season. Finally, we found that antisera collected from previously na�ve ferrets infected for the first time with influenza (that are commonly used for antigenic analyses during vaccine selection) do not accurately represent the Ab repertoires found in humans that have been infected or vaccinated multiple times with different influenza virus strains. We identified some individuals who have an Ab response targeted to a region of the HA of H1N1 viruses that recently acquired a mutation. Overall, our studies identify ways to improve the process of choosing seasonal influenza virus vaccine strains. We propose that the implementation of “sequence-first” surveillance, new cell culture systems, and the use of clinical human antisera for antigenic characterization of viruses will improve the process of selecting seasonal influenza vaccine strains.