A full understanding of the backscattering characteristics of wetlands is necessary for the analysis of the hydrological conditions. In this study, a temporal set of synthetic aperture radar (SAR) imagery, acquired at different frequencies, polarizations and incidence angles over the coastal wetlands of the Liaohe River Delta, China, were used to characterize seasonal variations in radar backscattering coefficient for reed marshes and rice fields. The combination of SAR backscattering intensity and an optical-based normalized difference vegetation index (NDVI) for long time series can provide additional insight into vegetation structural and its hydrological states. After identifying the factors that induce the backscattering and scattering mechanism changes, detailed analysis of L-band ALOS PALSAR interferometric SAR (InSAR) imagery was conducted to study water-level changes under different environmental conditions. In addition, ENVISAT altimetry was used to validate the accuracy of the water-level changes estimated using the InSAR technique—this is an effective tool instead of sparsely distributed gauge stations for the validation. Our study demonstrates that L-band SAR data with horizontal polarization is particularly suitable for the extraction of water-level changes in the study area; however, vertically-polarized C-band data may also be useful where the density of herbaceous vegetation is low at the initial stage. It is also shown that integrated analysis of the backscattering mechanism and interferometric characteristics using multi-mode SAR can considerably enhance the reliability of the water-level retrieval scheme and better capture the spatial distribution of hydrological patterns.