<p>This study investigates the global consensus problem of mixed-order multi-agent systems. Firstly, a control strategy based on switching functions is proposed to ensure the effective operation of fuzzy control in the fuzzy region. Secondly, the differential equation model satisfying the global Lipschitz condition is extended, and a more universal model framework is constructed. Furthermore, online learning laws with time-varying <i>σ</i>-modifications are introduced to effectively guarantee the global asymptotic convergence of the closed-loop multi-agent systems. Finally, the effectiveness of the proposed algorithm is validated through designed simulation experiments.</p>