Pan evaporation is the only long- term observation of potential evaporation around the world. In analyzing and predicting how droughts and hydrological cycles might change in a warming climate, change of pan evaporation is one crucial element to be understood. In this paper, we chose 416 sites with continuous monthly observations over 1960-2014 in China. We calibrated the wind function in the PenPan model to improve the estimation of pan evaporation. We developed a new approach, i.e., the experimental detrending (ED) approach, and made comparison with the traditional partial differential (PD) method in attributing changes of pan evaporation for the periods of 1960- 2014, 1960- 1993 (evaporation paradox) and 1993- 2014 (evaporation paradox disappeared). The results first showed that improvement in estimating pan evaporation can be made when using the new calibrated wind function: f_q(u_2)=3.977*10~(-8) (1+0.505u2). The comparison then showed that both methods can well attribute changes of pan evaporation, and the ED approach performs slightly better than the traditional PD method. In addition, the ED approach can help make effective adjustment for the PD method in attribution analysis so as to better understand the change of pan evaporation. Hence, the ED approach is recommended to assist a better understanding and prediction of water- energy cycles in a changing climate.