Frequent changes between pump and turbine operations pose significant challenges in the design of pump turbine runners with high efficiency and stability. In this study, a multiobjective optimization design system, including a 3D inverse design, computational fluid dynamics, design of experiment, response surface methodology, and multiobjective genetic algorithm, is introduced and applied to the design of a middle-high-head pump turbine runner. The key parameters in the design, including the blade loading, the blade lean at the high-pressure side of the runner, and the meridional channel shape, were selected as the optimized parameters. Two runners, one with a large positive blade lean and another with a large negative blade lean, were selected for further numerical investigations and measurements. Model tests show that both runners have good power performances. The runner with a negative blade lean has better stability than the runner with a positive blade lean. (C) 2015 Elsevier Ltd. All rights reserved.