Wide-area inundation caused by external flooding can result in catastrophic damage to lives and property. To minimize these impacts, it is essential to provide rapid and accurate information about flooding after its occurrence. Against this backdrop, this study proposes a method for high-speed and accurate flood prediction by combining a high-speed flood analysis technique utilizing Dynamic Mode Decomposition (DMD) with a correction method based on inundation sensor data. A flood prediction system was developed based on the proposed method, and its accuracy was verified using the 2015 Kinugawa River flood event. The results demonstrated that: (1) incorporating the correction method based on inundation sensor data improved flood prediction accuracy, (2) the high-speed flood analysis enabled rapid predictions, and (3) setting an appropriate number of DMD modes to avoid overfitting was critical for ensuring computational stability.