Pan evaporation, an indictor of potential evaporation, has decreased during the last several decades in many parts of the world; the trend is contrary to the expectation that the increase of actual evaporation will accompany global warming, known as the pan evaporation paradox. What is the essential relationship between pan evaporation and actual evaporation? This is still an uncertain problem. In this paper, the trends of pan evaporation and actual evaporation are investigated using observational data and observation-constrained simula-tion results using NCAR Community Land Model (CLM) in Xinjiang from 1960 to 2005. Our analysis suggests that the decreasing trend of annual pan evaporation accompanies the in-creasing trend of annual actual evaporation, the tendencies of them both have statistical sig-nificance (at 99% level and at 95% level, respectively). We also find that there is the same turning point in precipitation, pan evaporation and actual evaporation of 1986, and either before the point or after, pan evaporation has inverse trend comparing with actual evaporation and precipitation. The above analysis indicates that pan evaporation and actual evaporation have complementary relationship. These results support the issue of evaporation paradox described by Brutsaert and Parlange (1998) and suggest that decrease of pan evaporation indicates an increase of actual evaporation in Xinjiang in the past half century. The correlation analysis shows that diurnal temperature range (DTR), wind speed, low cloud cover and pre-cipitation are the most likely driving forces for the reduced pan evaporation and the ascending actual evaporation.