AbstractÀThis paper attempts to analyze and optimize the design of a liquid cooling module for an array of light emitting diodes (LEDs). Simulations of various models were carried out in FLUENT, CFD based software package with variations in five parameters to find the junction temperature of LEDs and pressure drop across the cooling module. The presence of pins in the cooling module showed a substantial reduction in the junction temperature, when compared to the case without pins, but the pressure drop increased a little. Optimization showed the capability of the liquid cooling module to dissipate as much as 180 W/cm 2 heat flux without exceeding the maximum allowable junction temperature. This paper develops a regression correlation as well as a neural network to compute the junction temperature of the LEDs for variations in five parameters. Further, a neural network is developed to find the pressure drop across the module. Optimization using a genetic algorithm was done both individually and combined for junction temperature and pressure drop.