<p>The multidimensional and multilayered propagation characteristics of public opinion in online social networks pose challenges to its prediction and control. To address this, hypergraph theory is introduced to characterize multidimensional interactions, and an improved susceptible-infected-hibernator-removed hypernetwork model (HA-SIHR) is proposed. Based on this, a multilayer social hypernetwork public opinion propagation model is constructed by incorporating multilayered network topology. The model simulates propagation processes driven by social reinforcement and weakening effects through the introduction of hyperedge thresholds and propagation time limits. The effectiveness and accuracy of the model are verified through simulation experiments and empirical validation. The results indicate that public opinion propagates more extensively in multilayer social hypernetworks coupled with scale-free hypernetworks and small-world hypernetworks. Furthermore, social effect mechanisms, propagation probability, and information credibility are found to significantly influence information propagation.</p>