This paper focuses on using confidence interval estimation while simulating a queueing system. The approach relies on the existence of regenerative points in the queueing process being simulated that produces a set of n independent observations. An application is given to a maintenance shop which is modeled as a multiserver queueing system. The shop is assumed to serve two classes of customers; passenger cars and cargo trucks. Comparison of results obtained from simulation using three different estimators is given. The approach can be applied to any queueing system that posses a recurrent steady state stochastic process.