The values of chilled water supply temperatures of chillers indicate their load distributions due to the chilled water return temperatures of all chillers being the same in a decoupled air-conditioning system. This study employs genetic algorithm (GA) to find out the chilled water supply temperatures of chillers for solving optimal chiller loading (OCL) problem. GA overcomes the flaw that Lagrangian method is not adaptable for solving OCL as the power consumption models or the kW-PLR (partial load ratio) curves include non-convex functions. This study uses the chilled water supply temperature as the variable to binary code chromosomes, and execute reproduction, crossover and mutation operation. After analysis and comparison of the case study, it has been concluded that this method not only solves the problem of Lagrangian method, but also produces results with high accuracy within a rapid timeframe. It can be perfectly applied to the operation of air-conditioning systems. Copyright (c) 2007 John Wiley & Sons, Ltd.