Recent clinical data on seasonal influenza mRNA vaccines have demonstrated suboptimal efficacy against the influenza B virus (IBV). We employed sequence optimization strategies that successfully enhanced the antigen expression of hemagglutinin (HA), and developed mRNA vaccine candidates targeting the WHOrecommended strains. When administered at a low dose (0.1 μg), both mono-and trivalent influenza A mRNA vaccines induced robust humoral immunity and conferred complete protection against homologous viral challenge in murine models, outperforming the quadrivalent inactivated vaccine (QIV, 2 μg). In contrast, IBV mRNA vaccines at an equivalent dose failed to elicit detectable antibodies and offered no protection, consistent with prior of suboptimal immunogenicity in human trials. These findings highlight strain-specific immunogenicity constraints inherent to conventional mRNA platforms. To overcome these limitations, we systematically compared three distinct RNA vaccine modalities: (1) nucleoside-modified mRNA, (2) self-amplifying RNA (saRNA), and (3) circular RNA (circRNA). Notably, a single 0.1?μg dose of trivalent saRNA vaccine elicited robust humoral immunity and provided complete protection against IBV challenge, whereas mRNA vaccination achieved only 14% survival. Importantly, long-term antibody monitoring over 20 weeks showed that saRNA at the low 0.1 μg dose maintained high antibody levels, with a markedly more durable response to IBV antigens than other platforms. Moreover, the trivalent mRNA vaccine exhibited a favorable safety profile, with no obvious abnormal body weight changes or serum biochemical abnormalities observed after immunization. Our findings advocate for strain-adaptive platform selection: conventional mRNA for generating rapid, high-magnitude responses against influenza A, and next-generation saRNA vaccine for enhanced dose efficiency, particularly against IBV.