Abstract: The littoral zone ecosystem of the Three Gorges Reservoir has been altered and degraded by prolonged winter flooding and summer drought. Taxodium ascendens was selected as an afforestation species for use in littoral woodland engineering. Sapling growth characteristics (e.g., height, diameter at breast height, crown width, foliage density and under crown height) were measured every year from 2009 to 2011. Statistical analysis of variations in these characteristics was conducted with SPSS 17.0 software, and redundancy analysis (RDA) was used to evaluate the growth responses of the saplings to influencing factors including growth characteristics and hydraulic factors (e.g., altitude, water-logging duration, submersion depth and duration of complete submergence). The results showed that the survival rate of T. ascendens saplings planted over 169 m above sea level was 84.52% after two rounds of winter flooding. Mean variations of height, crown width and foliage density were moderate, with values of 0.10 m, 0.04 m and -0.11% respectively. Mean variations of breast height and under-crown height were more obvious, with values of 0.78 cm and-0.64 m respectively. Rates of change of height, breast height, crown width and foliage density of T. ascendens under different submersion conditions were significantly different. Partially submerged saplings adapted well to the water level fluctuation, but all measured growth indices decreased as the flood depth increased. Up to 43.56% of the completely-submerged group and 11.76% of the partly-submerged group died back from the stem tip. Six influencing factors (altitude, water-logging duration, diameter at breast height, under crown height, foliage density and crown width) were identified by a manual forward selection method based on the Monte Carlo permutation test (499 permutations, P<0.05). RDA results indicated that 31.20% of sapling growth variations could be explained by the selected influencing factors, and the first two canonical axes accumulatively explained 82.90% of the species-environment relationships. Partial RDA results showed that 11.90% of sapling growth variation could be explained solely by the hydrological effects, and 14.40% by the initial growth characteristics effects; the combined explanation rate was 4.90%. It was concluded that littoral woodland engineering is a promising model for ecosystem restoration in the Three Gorges Reservoir given the potential ecosystem functions of providing habitat, acting as a carbon sink and intercepting pollution. It is proposed that future research may include:continuous monitoring, studies of understory vegetation and complex littoral woodland ecosystem construction and modeling. The results provide experiences and reference points for similar restoration projects in the littoral zone of the Three Gorges Reservoir.