The Multivariate Bayesian Regression model [41] is introduced that requires some knowledge of Multivariate as well as Bayesian Statistics. A concise introduction to this material was presented in Part I. This Chapter considers the unobservable sources (the speakers conversations in the cocktail party problem) to be observable or known. This is done so that the Bayesian Regression model may be introduced which will help lead to the Bayesian Source Separation models.