<p>In the digital media era, social platforms have evolved into a mainstream channel for information dissemination, profoundly reshaping the way public information flows and interacts. Topic heat has become a key indicator for constructing a scientific public opinion governance system. Considering the diversity of opinion leaders and public opinion information types, research of the factors influencing topic heat on online social platforms driven by host attributes and information types is of great significance for the construction of a modernized network governance system. In view of this, this paper collected 140,241 trending topics on Weibo from June 1, 2022, to May 31, 2023, used topic heat as explained variable. Regression analysis is employed to explore the association between host attributes, information types (political, cultural, social, economic) and topic heat, as well as their interaction effect on topic heat. Topic heat is measured as a composite index designed by integrating likes, comments and shares. Host attributes are classified into four categories: Ordinary users, Blue V(verified government/organizational accounts), Yellow V(verified individual accounts), Gold V(high-value content creators) based on Weibo’s certification system and influence level. The research shows that the contribution of host attributes to topic heat does not fully match their network influence rankings; in addition, the influence of different host attributes on topic heat varies significantly depending on the information type: Blue V(verified government/organizational accounts) is the best choice for hosts of political and social information; in cultural and economic information, Yellow V(verified individual accounts) and Gold V(high-value content creators) show unique potential for topic heat enhancement respectively. This finding extends existing theories. Theoretically, it advances the Two-Step flow of communication theory by revealing that influencer effectiveness relies on attribute-content matching. It also supplements the Uses and gratifications theory by refining the traditional framework with a ‘host attribute-information type’ matching mechanism. Overall, this study offers theoretical contributions by extending core communication theories and provides targeted practical guidance: for public institutions, it optimizes information dissemination strategies through attribute-content matching, which facilitates the diffusion of government affairs information; for social media platforms, it informs personalized content management and recommendation mechanisms, which also contributes to enhancing the visibility of enterprises engaged in social media information dissemination.</p>