Abstract: Abstract: In this thesis, collaborative optimization algorithm is applied to the multidisciplinary design optimization of axial-flow pump blades. Collaborative optimization algorithm being a kind of multidisciplinary design optimization has developed rapidly in recent years, and is popular among experts at home and abroad because of its special advantages. First, the main design variables which influence the hydraulic performance and structural strength simultaneously were referred to. Besides, the thesis described collaborative optimization algorithm that is used for complex engineering systems to optimize the design. The computing framework of collaborative optimization algorithm could also be found in this thesis. Moreover, the advantages and disadvantages of the collaborative optimization algorithm were also analyzed. In the part of the system-level mathematical model of collaborative optimization, the surface response method and constraint relaxing method were introduced, which changed equality constraint into inequality constraints. Because the constraint relaxing method needs less calculation and leads to higher rate of convergence, the thesis adopted the collaborative optimization algorithm. Then, the mathematical model for the axial-flow pump blades was set up with collaborative optimization algorithm. The system-level optimization model adopted multidisciplinary and multi-objective optimization method, i.e. the two disciplines-the hydraulic performance and structure-were expressed as an objective function of the system in the form of a linear combination. The subsystem optimization model should suffice the requirements of cavitations performance, cascade diffusion coefficient, and blades stresses. In the end, based on the iSIGHT optimization platform, hydraulic performance and structure were coordinated and optimized collaboratively. By a series of computational analyses, the model of axial-flow pump blades were exported and then tested. The experimental results show that axial-flow pump blades which are designed by this algorithm have a good comprehensive performance. In consequence, the case of the design optimization of axial-flow pump blades indicates that the convergence of the collaborative optimization algorithm based on constraint relaxing method is reliable, and solved the problem that coupling multiple disciplines would have large volumes and complex data. At the same time, it verified the feasibility of a multi-disciplinary design optimization in the pump blade design optimization field. As to an axial-flow pump, collaborative optimization algorithm does produce a better-optimized design scheme, improve the overall performance of axial-flow pump blades, and extend the range of application of the axial-flow pump blades. Meanwhile, this algorithm avoids controlling the related indexes by virtue of mere experience when the design optimization is considered in a single discipline. Therefore, the collaborative optimization algorithm for axial flow-pump blades of multidisciplinary design optimization is efficient and feasible.