Codes and standards specify minimum performance of HVAC equipment such as water-chilling packages commonly known as chillers. Chiller performance is usually described using two metrics, a full and part-load efficiency. Chiller operation is modeled in building energy simulation software, using a set of performance curves that describe the performance of the equipment as environmental and operational conditions change. While code and standards define the minimum equipment efficiency, simulation-ready performance curves meeting these ever-changing requirements are not typically available. This paper discusses the use of a genetic algorithm to modify typical chiller performance curves to match both full and part-load code minimum efficiency.