To address the issues of structural differences in public knowledge, difficulty in quantifying its value, uncertainty and ambiguity in expert preferences, and challenges in measuring expert reliability, this paper proposes a large-group emergency decision-making method that integrates public relational networks with expert preference reliability. First, a public relationship network is constructed based on the similarity of user viewpoints. A public knowledge value measurement model for emergency decision-making is then developed by integrating two factors: the informational value of user-generated content and the relational value derived from user connections, thereby completing the valuation of public knowledge. Simultaneously, to address the counterintuitive outcomes of existing probabilistic linguistic term set reliability formulas, a novel reliability measurement method is introduced. This method comprehensively accounts for the reliability of expert preference information and the trust relationships among experts, enabling the measurement of expert reliability under varying attributes and the determination of corresponding expert weights. Next, a fusion coefficient is calculated by balancing attribute ranking consistency and differences in attribute importance. This coefficient integrates attribute weights derived from both public knowledge and expert reliability to obtain comprehensive attribute weights and uses aggregation operators for information aggregation and scheme ranking. Finally, the feasibility of the method was verified by taking the torrential rain in Hebei Province as an emergency decision-making case, and the stability and superiority of this method were verified through sensitivity analysis and comparative analysis.