The wetland restoration process is usually disturbed by the undesirable impacts of invasive plants and ruderal wetland plants. Soil seed banks are usually useful indicators of the efficiency of wetland restoration. The Momoge National Nature Reserve (MNNR) is the largest wetland reserve in Jilin province, China and was established mainly for protecting wetland habitats and rare bird species, including the Siberian white crane (Grus leucogeranus). Tubers of Scirpus planiculmis Fr. Schmidt are the main food source for the Siberian white crane. However, the Scirpus marshes in MNNR have been degraded severely by global climate changes and human activities. Since 2010, the local governement has rewetted many degraded marshes to restore the habitat of the Siberian white crane. In the restored marshes, the dominant species of plant was Echinochloa caudata Roshev., a native ruderal species, which often exhibited over 70% coverage. However, the target plant S. planiculmis has rarely been found in the restored wetlands. In this study, we compared the seed abundances of S. planiculmis and E. caudata in soil seed banks. We also examined the growth and seed setting characteristics for both species. The seed density of S. planiculmis in the soils of natural wetlands was 18.42 times higher than that of E. caudata, while the seed density of E. caudata in restored wetlands was 4.04 times higher than that of S. planiculmis. The high seed productivity of S. planiculmis (9210.4 ± 1513.4) seeds/m² and the relative higher buoyancy of their seeds (FP₅₀ = 39.7 days) supported the potential for restoration via the soil seed bank. Moreover, the density of active seeds in restored wetlands still reached (664.32 ± 105.98)seeds/m², which was still significant lower than that in natural wetlands. The dormant soil seed bank of E. caudata in invaded wetlands (3345.9 ± 520.3)seeds/m² was significantly greater than that in natural wetlands, which indicated that the restored wetlands were heavily encroached by E. caudata. Additionally, the high seed production (7621.4 ± 376.25)seeds/m² and lower seed buoyancy (FP₀ = 5 days) of E. caudata indicated that E. caudata could not expand efficiently through water dispersal. Furthermore, both plant height and vegetation cover of E. caudata were significantly higher than those of S. planiculmis ((110.83±17.57) versus (68.33±10.49) and (75±4.35) versus (30±6.24), respectively), which could affect water dispersal of the buoyant seeds and influence species expansion. The MNNR wetlands, which are located in a mid-temperate zone with long winter days, reduced the decomposition of Echinochloa litter. Additionally, reproductive production of S. planiculmis was significantly reduced because of the increase in shade. Thus, improving the availability of light for S. planiculmis and diminishing the seed bank of Echinochloa by removing live stands and litter before the flowering phase may be an efficient strategy for restoration of S. planiculmis. More studies on the responses of soil seed banks of S. planiculmis to E. caudata removal are needed.