The altered water regime in the Three Gorges Reservoir of the Yangtze River, China, has caused many serious environmental problems. Among them, cadmium pollution is serious, so vegetation restoration and phytoremediation of Cd-contaminated soil is fundamental in this region. The soil moisture regime of the soil is an important factor for the physical, chemical, and biological properties of soil. The bioavailability of heavy metals will be reduced under environmental flooding, making it difficult to remove the heavy metals during flooding. Research has shown that Salix variegata can tolerate both flooding and heavy metal stress. Therefore, S. variegata could be a suitable species to restore the vegetation and remedy Cd-contaminated soil in the Three Gorges Reservoir Region. To explore the phytoremediation of Cd-contaminated soil by S. variegata under flooding, factorial experimental treatments included two water regimes and four Cd concentrations. The water treatments included ambient water supply (CK) and light flooding (FL). The cadmium additions included control (0 mg/kg), low concentration (0.5 mg/kg), middle concentration (2 mg/kg), and high concentration (10 mg/kg), and each treatment involved a group of planted S. variegata shoots and a reference group without plants. Soil and water samples were collected 30 and 60 days after the start of the treatments. A sequential extraction was used in this research. The weakly bound fractions of Cd (exchangeable and carbonate-bound) were considered easy to mobilize, so they possibly presented a potential risk to the environment. The mobility factor (MF) was used to describe Cd behavior in soil. The results showed that: (1) With increasing Cd concentration in the soil, the labile fractions of Cd increased significantly, meanwhile the Cd bioavailability increased significantly, and the environmental risk could be increased significantly; (2) Water treatment significantly affected the speciation of Cd in the soil, and the MF values under the flooding treatment were much less than those under the ambient water treatment at days 30 and 60 of the experiment (P < 0.05). Furthermore, the phytoremediation of Cd-contaminated soil under environmental flooding was very different from that of the soil under ambient water supply. (3) S. variegata did not affect the total Cd concentration or Cd speciation of the soil samples taken on day 30 of the experiment, whereas it significantly reduced the total Cd concentration of the soil day 60 under the FL treatment. The exchangeable and carbonate-bound Cd concentrations of soil from both the CK and FL treatments on day 60 were also significantly affected by S. variegata. (4) The MF value of the soil samples was significantly reduced by S. variegata compared with the non-planted groups, and the average decline was 2.0%, 4.12% in treatments of PCK and PFL on day 30 of the experiment, and 9.71%, 9.32% in treatments of PCK and PFL on day 60 of the experiment, respectively. There are two primary reasons explaining why the total Cd concentration of the soil samples was not considerably affected by S. variegata shoots: (1) The duration of the incubation experiment was too short; (2) the biomass of the S. variegata shoots was relatively small. However, S. variegata significantly reduced the MF value both in the ambient water and flooding treatment, but did not differ significantly between both treatments. In conclusion, S. variegata could be a suitable species for the phytoremediation of Cd-contaminated areas under environmental flooding.