Every musician who improvises has a unique musical style or musical vocabulary which he or she uses in creating music. This may be perceived in his or her compositions. This thesis is the design of a computational model for predicting the improvisational decisions made by a particular blues musician, Stevie Ray Vaughan, for a given input scenario. It brings into effect the fact that creative works involve the use of pre-existing structures stored in the creator's mind or knowledge base, retrieved and reconstructed on the basis of appropriate rules, which are triggered by the nature of specific input. These structures or musical patterns were deciphered from selected works of the musician by a process of reverse engineering. The input parameters contributing to the use of these patterns were also extracted. The computational model was partially implemented as a limited production system. A probabilistic method involving input parameters and musical patterns was used to generate prediction rules. This partial implementation was tested with three different input scenarios. It was found that the model achieved its goal of predicting the musician's decisions with a limited degree of accuracy. More significantly, the tests provided valuable insight into ways to improve the performance of the existing model. One of the important recommendations is the revision of the definition of a pattern to include limits on its duration.