Pan evaporation decrease has been reported worldwide over them past decades. A recovery trend, even an increasing pan evaporation trend, has been recently found. Remarkably, most studies on Chinese pan evaporation change in China were based on simulations involving meteorological variables, including temperature, radiation (sunshine duration), wind speed and relative humidity, due to the pan evaporation observation inconsistency caused by the micropan (D20) replacement with large pans (E601) around 2002. In addition, it has been reported that a large-scale humidity sensor replacement across China has occurred since the 2000s, which can cause an underestimation of relative humidity and in turn leads to an inconsistency in simulated pan evaporation. Therefore, the recent pan evaporation trend independent of the observed relative humidity in China must be revisited. In this study, we complete the D20 pan evaporation from 1988 to 2017 according to E601 observations under the constant conversion coefficient assumption between the evaporation observations of these pans in the same month of every year at each station and conduct trend and attribution analysis through linear regression and PenPan-D20 model partial differential methods, respectively. A significant 2.68 mm/a/a upward pan evaporation trend (P<0.05) from 1988-2017 is revealed, primarily driven by the air temperature rise across China. Humidity sensor replacement causes an ~1.3% relative humidity underestimation, producing nonnegligible pan evaporation trend simulation errors.