Pumped storage unit (PSU) operated under the combination of different condition, which may cause equipment wear, fatigue and fault issues. It is difficult to achieve accurate and stable degradation trend prediction (DTP) considering only either the accuracy or stability. In this paper, a combined multi-objective optimization model is proposed for DTP of PSU. Firstly, the health model of PSU is constructed with back propagation neural network (BPNN) and condition parameters of active power, working head and guide vane opening. Subsequently, the performance degradation index (PDI) is developed to characterize degradation level of PSU. Then, considering high accuracy and strong stability in predicted model, kernel extreme learning machine (KELM) is optimized by multi-objective particle swarm optimization (MOPSO) to achieve DTP. To validate the effectiveness of the proposed model, the monitoring data collected from PSU in China is taken as case studies. The results demonstrate that the proposed model outperforms other comparison models in terms of predicting accuracy and stability.