HW/SW techniques make it possible for the system designers to validate their design, assign modules to be implemented in either hardware or software in the early stages of the system design life cycle. In addition, those techniques provide powerful mechanism for continuous system validation until the final product is done. Partitioning the system into either hardware or software, in the system early stages, is vital decision that has to be done iteratively and accurately. Many techniques have been proposed for HW/SW partitioning: conventional circuit partitioning techniques, simulated annealing, expert systems, and even genetic algorithm techniques. The partitioning problem has been proved to be and NP-Hard problem, thus AI, ANN and GA techniques can find a rich playground to apply their techniques. This paper presents a novel approach to use Bayesian Belief Networks as the tool that does the partitioning decision when provided by simulation parameters that measure certain characteristics in the design.