Nanomaterials, because of their unique properties such as extremely small size and increased ratio of surface area to volume, have a great potential in many industrial applications that involve electronics, sensors, solar cells, super-strong materials, coatings, drug delivery, and nanomedicine. They have the potential also to improve the environment by direct applications of these materials to detect, prevent and remove pollutants. While nanomaterials present seemingly limitless possibilities, they bring with them new challenges. Among them is the precise control of the morphology of nanomaterials, which is extremely critical to the development of advanced nanodevices with various functionalities. The one-dimensional nanostructures of Cadmium Selenide (CdSe) have been found to represent morphologies of nanowires, nanobelts, and nanosaws, however, their synthesis is by trial and error. Predictive modeling and control methods are essential to process yield and productivity improvement. The process yield (response) is a vector whose elements correspond to the number of appearances of the different types of nanostructures, namely nanosaws, nanowires, and nanobelts. The goal in this paper is to apply existing Bayesian methodologies to describe the growths of these nanostructures in terms of process variables and to predict the probability of transition from one nanostructure to another when changes are made to one or more process variables. We also propose a Bayesian algorithm to identify the optimal process conditions that maximize the predicted probability of each type of nanostructure.