This report describes the development of a numerical optimization procedure for the analysis of marine propellers. The activity is performed by INSEAN in the framework of WP35 of the STREAMLINE Project and the present report is written in fulfilment of Deliverable D35.4. The work described in the report is aimed at developing and integrating numerical optimization tools to build an optimal design framework for the improvement of state-of-art screw propellers in real operating conditions. The optimization problem is recast as a shape optimization study and computational strategies to achieve an efficient and flexible parametrization of propeller blade geometry are reviewed. Classical Free Form Deformation techniques are used as a reference to develop a novel approach, called Conformal Free Form Deformation. The capability of this technique to perform propeller blade shape manipulations is demonstrated through examples. Alternative optimization frameworks are outlined and compared. The importance of introducing a Robust Design Optimization model is clarified in order to make possible the analysis of propeller performance over a representative range of variations of operating conditions instead of considering an isolated design point. In addition to that, different algorithms based on local optimization and global optimization strategies are compared. The impact of introducing design constraints into the procedure is then discussed through simple example describing propeller shape manipulation studies based on a limited number of parameters. Requirements for the selection of a propeller hydrodynamics model to be interfaced with the optimization model are analysed. The Boundary Element Method developed as part of WP34 activities is chosen as an adequate trade-off between computational efficiency and accuracy of numerical predictions of propeller performance. Finally, numerical examples are presented and results of simple propeller optimization studies are discussed to analyse the capabilities of the proposed methodology.