During public emergencies, the influx of patients leads to a decrease in intensive care unit(ICU) care efficiency. To comprehensively evaluate the effectiveness of patient bed allocation strategies in improving ICU care efficiency, an agent-based modeling-discrete event simulation(ABM-DES) method was used to construct a simulation model of the ICU care system. Taking the comprehensive ICU in Renmin Hospital of Wuhan University as an example, four differentiated bed allocation schemes were designed based on the zoning management strategy of the emergency department. An evaluation system for ICU care efficiency was constructed based on the entropy-weighted technique for order preference by similarity to an ideal solution(TOPSIS) method, incorporating multiple dimensions including patients, medical staff, supplies, and space. The care efficiency under four patient arrival modes was quantitatively assessed using this system. The optimal bed allocation scheme was subsequently identified through comparative analysis. The results show that patient arrival modes significantly influence the optimal bed allocation scheme. Furthermore, the efficiency enhancement effect of the zoning management strategy manifests exclusively in the long-term high-volume arrival mode. The findings provide critical guidance for ICU response to public emergencies.