Flash floods are considered as the most intense hazard therefore rapid identification is needed. Tsunami also causes flash floods as it is commonly generated around the Pacific Ocean. Flash floods are also caused by the severely blocked streams in heavy rainfall. Floods have ended up
so many lives more than the other natural hazards and also devastated precious belongings and infrastructures. Cattles have also been affected by the floods event. Floods devastate the construction and infrastructure like roads, bridges and buildings that comes in the vicinity of effected
area by flood. Breakdown and overflow of dams may produce the deadly flash floods to the populated area and environs. Many strategies and methods have been followed to determine the flash floods on early basis so that evacuation announcements may be propagated. Numerous researches have been
studied and carried out to accomplish this task. Development of dams and reservoirs have been given more significance. Artificial Intelligence based competent decision-making algorithms like Bayesian classifier, PSO, ANN, NNARX, SVM and GA have been applied to achieve more accuracy in predictive
analysis. Direct observations from the sensors and data from the meteorological department have also been used for the predictive analysis of flash floods. Many yardstick parameters have been proposed in past researches to identify the flash floods vigorously like environmental CO2 levels,
precipitation velocity, wind speed, upstream level, height of the water, pressure, temperature and cloud to ground flashes. In this research papers a novel Artificial Intelligence based approach Modified Cuckoo Search (MCS) has been adopted to forecast the flash floods more rapidly and accurately.
Obtained results in the MATLAB have proved that Modified cuckoo search with the combination of Artificial Neural Network worked better than the recent available methods. Results have also been validated by comparing the MLP-PSO.