<p>Negative public opinion on the internet in emergencies induce emotional fluctuations in decision-makers. How negative public opinion in emergencies acts as a stressor affecting emergency decision-making, and how intervention strategies can mitigate its adverse effects, is a research topic worthy of attention. Based on this, an emotion perception-based decision-making method for emotional intervention is proposed. First, mining social media data (using SnowNLP) to quantify negative opinion increments as stressors, linking them to decision-makers’ emotions. Second, based on the characteristics of decision-makers’ emotional states, a large-scale group decision-making preference evolution model is constructed, taking into account psychological differences in emotional states under emotional black-box scenarios. Third, an emotion-perception model is constructed to classify decision-makers by emotional stability and establishing an intervention-based preference evolution model. Depending on different decision-making needs and emotional states, a decision preference evolution model under emotional intervention is established. Finally, using Typhoon “Yagi” that occurred in Hainan, China, in 2024 as a case study, the comparative analysis was performed to examine the differences in the evolution trends of large-group decision preferences before and after emotional intervention, verifying the necessity of emotional intervention strategies in emergency decision-making.</p>