There is increasing concern about the sustainability of large-scale plantations in water-limited regions, with most studies indicating that the higher water use of forest plantations compared to herbaceous crops exacerbates water shortages in such areas. However, the mechanisms that control evapotranspiration in forest plantations under complex water, energy, and climatic conditions and across diverse geographical regions remain unclear. Here, we report continuous water flux data for a poplar plantation (Populus×euramericana ‘74/76’) in Daxing District, Beijing, China, collected with an eddy covariance (EC) and microclimate monitoring system. Our objectives were to a) quantify the dynamics of evapotranspiration (ET), surface resistance (R_s), and climatological resistance (R_i) over dry and wet years; b) examine the direct controlling effects of biological and environmental factors on ET by using partial correlation analysis under different soil moisture conditions over dry and wet years. On the interannual scale, average daily ET in dry years (2.23±1.30) mm/d was 17% lower than that in wet years; surface resistance (R_s:LAI) increased by 50% in dry years, but there were no significant differences in R_i and R_a between dry and wet years. At the seasonal scale, seasonal drought stress had a discernible impact on ET, R_s, and R_i of the poplar plantation, and water supply (precipitation+irrigation) caused 71% of seasonal variation in ET (P<0.01). Partial correlation analysis indicated that R_s was the main factor controlling daily ET, except under severe water stress (REW<0.1), and daily ET was negatively related to R_s (second-order correlation coefficient of -0.518 to -0.293, P<0.01). The effect of R_s on ET was stronger in dry years than in wet years, and the effects of R_i and R_a on daily ET were irregular and weaker than that of R_s. Daily ET of the poplar plantation was directly affected by relative extractable soil water (REW) only under water stress (REW<0.4) in dry years (Pearson coefficient 0.217-0.323, P<0.01), and it was indirectly influenced by REW under other soil water conditions. Compared to partial correlation analysis, correlation analysis would incorrectly evaluate the effects of R_s, R_i, and R_a on ET.