Abstract: ‍ ‍River level monitoring has been widely used to ensure water hinge safety and warn disasters caused by floods. The traditional monitoring method is realized by setting a water level meter on site， which is difficult to implement in remote areas with inconvenient transportation. To solve this problem， a remote sensing monitoring method for river level based on time series SAR images is proposed in this paper. Based on bridge multiple scattering mechanism， by extracting the position changes of bridge multiple scattering signals in time series SAR images， the river level change monitoring under the bridge can be realized. In addition， in order to suppress the interference of ship strong scattered signals， a bridge energy accumulation algorithm is proposed. Firstly， the pre-processing operations such as co-registration and median filtering are carried out on the time series SAR images， and the cascade fitting processing based on morphology is used to accurately extract the direction of the bridge in the SAR amplitude images， then the migration of the bridge signals in the azimuth-direction is corrected. Finally， energy accumulation is carried out on the scattered signal of the bridge along the azimuth-direction， and the scattered signal of the ship is dispersed due to the inconsistency with the direction of the bridge. This process greatly improves the signal-to-clutter-plus-noise ratio of the scattered signal of the bridge， so as to ensure the accurate extraction of the bridge multiple scattering signals. The accuracy of the proposed method is analyzed from two aspects： imaging defocusing and extraction error of the double-bounce signal in SAR images. The on-orbit SAR satellites can reach the accuracy of water level monitoring at the decimeter level due to the limitation of resolution. In the future， with the improvement of SAR image resolution， the proposed method is expected to achieve water level monitoring accuracy of several centimeters. The COSMO-SkyMed and Sentinel-1A data are processed to validate the feasibility of the proposed method. By comparing the results obtained by the on-site water level meter， the proposed method can achieve estimation accuracy better than 0.38 m using the existing spaceborne SAR data， such as COSMO-SkyMed datasets.