During respiratory demand peaks, such as seasonal influenza outbreaks or COVID-19 surges, healthcare systems often face significant strain, especially in Intensive Care Units (ICUs). Bed shortages and long waiting times can lead to delayed care and worsened patient outcomes. To address this, healthcare systems increasingly turn to digital technologies, such as digital twins, to optimise patient flow and resource allocation. This paper illustrates the implementation of digital twins for managing bed waiting times in intensive care units during respiratory demand peaks. First, we described the patient’s journey from the ED to the ICU using Supplier-Input-Process-Output-Customer (SIPOC) diagrams. After this, we analyzed input data analysis, verifying the process variable data’s randomness, heterogeneity, and goodness-of-fit. We then modelled the ED through a digital twin designed in ARENA® software. Following this, we validated the model by applying a Kruskal Wallis test on the waiting time for ICU beds. Lastly, we pretested two improvement scenarios: increasing the number of ICU beds by i) 3 and ii) 5. The suggested method was applied in a European hospital group during one of the first COVID-19 waves. The outcomes revealed that the waiting time for ICU beds (1.88 h) can be meaningfully reduced if strategy ii) is applied.