Light use efficiency (LUE) is a major limiting factor of gross ecosystem productivity (GPP). Various LUE models have been established to evaluate regional GPP. The various maximum light use efficiency (LUE_max) values used in these models are critical variables that influence model uncertainty. Since the dynamics and influential factors affecting LUE at different temporal resolutions vary, it is unclear whether the values of LUE_max at these temporal resolutions differ. Therefore, we examined the dynamics of LUE and LUE_max using data from a poplar plantation (Populus euramericana cv.) in the Daxing district, Beijing. Eddy covariance measurements were taken at this study site. A multiple stepwise regression procedure and recursive partitioning methods were applied at both monthly and annual scales. The results indicate that the averagely daily LUE values from 2006 to 2009 were (0.33 ± 0.16) gC/MJ, (0.35 ± 0.23) gC/MJ, (0.39 ± 0.16) gC/MJ, and (0.32 ± 0.19) gC/MJ, respectively. The daily LUE varied seasonally, with a rapid increase occurring in April and May, a peak from Jun to Aug, and a gradual decrease after September. The factors influencing daily LUE were different during different parts of the growing season. Air temperature (Ta), evaporative fraction (EF), and vapor pressure deficit (VPD) were the main factors in affecting LUE in April. In May, photosynthetically active radiation (PAR), volumetric water content (VWC), EF, and canopy conductance (g_c) were the factors with the greatest influence. PAR, VWC, g_c, and VPD had large impacts on LUE in June. In July and August, LUE was controlled by PAR and g_c. In September, PAR, soil temperature (Ts), VWC, and EF were the main influencing factors, while PAR, VWC, EF, g_c, and VPD influenced LUE in October. PAR was the most important factor regularizing LUE in the middle of the growing season, while moisture conditions were the main influencing factors early and late in the growing season. However, monthly PAR (PAR_m) was not a main factor affecting monthly LUE (LUE_m). In contrast, 71% of LUE_m variations were explained by the monthly evaporative fraction (EF_m) and monthly soil temperature (Ts_m). Because of various influential factors, LUE_max were not identical among temporal resolutions. Recursive partitioning analysis showed that EF=0.42 was the node for LUE in April. Correspondingly, LUE_max in Apr was 0.22 gC/MJ, when EF ≥ 0.42. PAR and EF were the nodes for LUE in May, LUE_max in May was 0.39 gC/MJ, when 17 ≤ PAR < 27 MJ and EF ≥ 0.77. In June, LUE_max was 0.38 gC/MJ when VPD < 1.2 kPa and PAR ≥ 21 MJ. From July to October, PAR was the main node for LUE, when LUE_max was 0.66 gC/MJ, 0.69 gC/MJ, 0.61 gC/MJ, and 0.44 gC/MJ, respectively. LUE_max in July, August, and September was slightly larger than that in other months. The average annual LUE_max was approximately 0.44. We concluded that iLUE models should incorporate different LUE_max at different temporal scales to better model GPP.