A multi-surrogate based optimization is a technique for the optimization of a turbomachinery impeller shape, which reduces the total design time. Computational analysis and surrogate-based optimization methodology have been implemented for the shape optimization of a centrifugal impeller blade in this problem. Geometric parameters such as impeller inlet and two-point control Bezier curve blade profile were selected as design variables. The objective was to maximize the total head as well as the efficiency of the pump using surrogate based optimization methods. A sample space was created for the design variables based on the sensitivity of the variable to the objective function and literature survey give a range of sample space. A three-dimensional Reynoldsaveraged Navier-Stokes equations were solved to analyze the performance of the pump and the results were validated with experimental results. The simulations were done at different design points within the sample space to train the surrogate and to find the design optimum point using genetic algorithm. The shape of the impeller was optimized, to decrease the input power and improve the total head by varying the inlet and exit angles of the impeller. Copyright © 2016 by ASME