Abstract: In recent years, flood disasters have occurred frequently and serious impacts have been brought to society. During flood disasters, significant changes in river water levels and precipitable water vapor (PWV) are often accompanied. Taking the flood disaster in Porto Alegre, Brazil in 2024 as an example, data from GNSS stations were selected, and research on flood water level and water vapor monitoring was conducted. The results indicate that, before the rainstorm, the correlation coefficient between the water level inversion at the SPH4 station and the data from the hydrological station is 0.993, and the root mean square error (RMSE) is 0.02 m. During the rainstorm, the correlation coefficient between the water level inversion results of the SPH4 station and the IDP1 station on both sides of the river reaches 0.997, and the RMSE is 0.06 m. There is a time difference of 2 to 5 days between the rainfall peak and the water level peak. The correlation coefficient between the PWV value inverted by the GNSS station and the PWV value measured by the radiosonde station is 0.992, and the RMSE is only 1.9 mm. The maximum rainfall occurs within 5 h when the PWV value reaches its peak. The experiment demonstrates that shore-based GNSS equipment can accurately invert the changes in flood water levels and precipitable water vapor, and broad application prospects are held in the prevention and monitoring of flood disasters.