This paper suggests the development of a novel cold plate that is predicated on a mesh channel and performs multi-objective optimization with parameters such as coolant flow direction and velocity, cold plate thickness, inlet length-to-width ratio, and structural fillet to balance the maximum temperature (Tmax), temperature difference (ΔT), pressure drop (ΔP), and mass (m), identifying the optimal cold plate structure. The optimal design is then compared with other reconstructed cooling structures to demonstrate its superiority. The results show that the designed optimal structure reduces Tmax, ΔT, ΔP, and m by 9.7%, 22.2%, 77.3% and 8.9%, respectively, compared to other cooling structures. Finally, the accuracy of the numerical simulation is verified experimentally with an error of less than 3%.