Simulation has long been used as a decision making tool in manufacturing environment. It enables management to have a better understanding of the behavior of the production facilities. It also allows the management to predict the effects of varying certain production parameters on the whole production system. To convince the user of the validity of the simulation results, accurate modeling and simulation of the physical system is necessary. This requires the various fine details of the factory environment to be modeled in the simulation, resulting in a large number of events to be generated and processed. Simulating these events is known to be a long process and in some cases the simulation can take days to complete. However, simulation must deliver results in an acceptable time-frame for swift decisions to be made. Parallel discrete event simulation provides two advantages over its sequential counterpart. Firstly, the combined computing power of the parallel processors promises shorter execution time for the simulation. Secondly, the user can run a larger model on a parallel system with larger memory. The aim of this project is to study the feasibility of using parallel discrete event simulation in the modeling and simulation of a complex manufacturng system such as a Virtual Factory. Extensive literature survey is carried out to determine the current state of the art technology in the field of parallel discrete event simulation and to examine the work done by existing research groups working in this area. A simulation prototype has also been developed to experiment with different factors that may affect the use of parallel discrete event simulation in the modeling and simulation of a virtual factory. A conservative superstep protocol has been implemented. Successive refinements have been made on the original protocol. Experiments have been carried out and some performance results are reported in this thesis. Master of Applied Science