Tremendous efforts had been given to ensure proper heat dissipation in electronics cooling but very seldom consider design robustness and user preferences in design principles of the heat-dissipating devices. Multi-objective optimization problems are one of the preferences elicitation tools that could be used and is highly visual on the costs and benefits associated in choosing different preferences. It would be better if a wider temperature range is offered for thermal management schemes and is made available if the user desires. It would also be sought upon if automatic determination of the user preference for a wider range of varying performance were available. In this paper, a liquid impinging heat exchanger with a thermoelectric module was chosen as the example of how this paradigmatic scheme was implemented using black box models. An orthogonal sampling method was applied with three parameters considered. The temperature at the interface between the chip surface and the liquid impinging thermoelectric cooler (LITEC) is taken as the desired response. A response surface was generated using Kriging method, after which, a multi-objective optimization problem was then formulated to include robust definition and user preference for energy efficiency. The optimal operation parameters of the inlet flow velocity and the thermoelectric (TE) chip control voltage were found for various levels of heat loading conditions and different considerations of design robustness and energy awareness.