Water level information is one of the basic parameters to ensure the safety of reservoir dams and optimize the operation of water resources. In recent years, the research on monitoring altitude difference (snow depth, sea level) and surface environment (vegetation, soil moisture, temperature, and so on) using Signal-to-Noise Ratio (SNR) of GNSS observation data has become a new research hotspot. This paper analyzes the characteristics of the GNSS SNR data which is in the case of low satellite elevation angles, and the inversion principle of GNSS-MR technology based on the SNR data of single geodetic-quality Global Navigation Satellite System (GNSS) receiver to detect water level is given in details. The water level inverted by using the deformation monitoring network data of Shuangwangcheng Reservoir, an important reservoir in the East Route of South-to-North Water Transfer Project, was compared to water level data from nearby and in situ sensor. The result indicates that the variation daily time series of water level estimated by GPS L1 band SNR measurements can be achieved to 4.7cm, 4.9cm and 3.8cm, compared with the measured daily water level, the correlation between the estimated value and the daily records can even reach up to 0.99. The accuracy of inversion results based on GPS L2 band data is lower than L1, and the single station RMS using L2 data is about 8.9 cm. What's more, reservoir water level inversion method based on GNSS SNR can be used as an effective supplement for reservoir water level monitoring system. For Beidou system, the SNR data of B1 signal is used for water level inversion. Because of the number of observable satellites, the accuracy is worse than GPS. The RMS, compared with the measured data of water level station, is 6.1 cm. The preliminary study shows that the GNSS-MR water level inversion method can provide real-time water level information and can be an effective supplement to the reservoir water level monitoring system.