With smaller inlet flow velocity, a micro-channel stack requires less pumping power to remove a certain rate of heat than a single-layered micro-channel, because it provides a larger heat transfer area. A simple thermal resistance network model was developed to evaluate the overall thermal performance of a stacked micro-channel heat sink. Based on this simple model, in this study, a single objective minimization of overall thermal resistance is carried out using genetic algorithms. The aspect ratio, fin thickness and the ratio of channel width to fin thickness are the variables to be optimized, subject to constraints of maximum pressure drop (4 bar) and maximum volumetric flow rate (1000 ml/min). During the optimization, the overall dimensions, number of layers and pumping power (product of pressure drop and flow rate) are fixed. The study indicates that reduction in thermal resistance can be achieved by optimizing the channel configuration. The effects of number of layers in the stack, pumping power per unit area, and the channel length are also investigated.