The ongoing COVID-19 pandemic, intensified by emerging SARS-CoV-2 mutations, highlights the urgent need for enhanced vaccines. Despite considerable efforts in vaccine design, improvements are still required in formulating vaccines targeting the novel coronavirus. This study, utilized immunoinformatics and reverse vaccinology to design multi-epitope vaccines targeting emerging variations. B and T cell epitopes were generated by analyzing the mutation sites of the prevalent variant strains, and two vaccines were designed by linking with two different adjuvants. Interaction of the model vaccines with four Toll-like receptors (TLR) revealed a relatively high affinity between vaccines and immune receptors. Codon optimization and computational cloning were conducted to validate the robustness of the multi-epitope vaccines and immunogenic simulations were performed to assess the antigenicity and antibody generation capability of the vaccine. The L455S mutation in the JN.1 variant and its adjacent F456L mutation on antibody effectiveness against the XBB variant revealed that 15 antibody structures maintained a certain level of binding affinity. This study offers an immunological evaluation from a mutation-centric perspective and integrates co-evolutionary analysis with immunoinformatics to design effective multi-epitope vaccines targeting various SARS-CoV-2 strains. The methodologies applied in this research can also be extended to the vaccine development for other pathogens.