This paper presents a CFD-based, heuristic evolutionary algorithm for shape design and optimization of baffled fluid distributor. In this algorithm, the baffle surface is firstly divided into numerous identical control areas (volumes), each control area having an orifice in the middle. Under the constraint of constant global porosity of the baffle, the algorithm adjusts the size distribution of orifices so as to approach identical mass flowrate passing through every control area. An automatic program is processed iteratively so that the baffle configuration evolves toward the optimized shape, providing a uniform flow distribution among parallel outlet channels. To illustrate the principles and procedure of this algorithm, a 2D example of baffled fluid distributor is introduced and tested. Numerical results show that this algorithm can successfully reach uniform flow distribution with a small pressure drop increase. Sensitivity analysis also shows that this algorithm is robust, effective, general and flexible compared to traditional arbitrary or empirical propositions. A parametric study of influencing design parameters on the performance of the algorithm is carried out in order to provide some design guidelines. Finally, the theoretical basis and essential steps for the extension of the current algorithm to tackle 3D problem are established. The easy implementation of this simply solution for the general fluid maldistribution problem demonstrates its promising application in real engineering field. (C) 2014 Elsevier Ltd. All rights reserved.