In this paper, we describe the capacity management of Intensive Care Units (ICUs) in a 300-bed multi-specialty hospital where the alternative ICU is utilized when the appropriate ICU is full for a set of two types of ICUs. Inter-arrival time and service time distributions in these ICUs have been tested and found to be exponentially distributed. While most capacity management models are deterministic in nature, we have developed a queuing model to provide a basis for decision-making in the design and management of these ICUs. The model results in around 19800 linear steady state equations, which are solved using the CPLEX linear optimization solver. Based on real data available from a hospital in India, the results demonstrate that the utilization of the ICU beds will improve up to 28 percent when admissions to the alternative ICU are permitted.