Damage caused by floods in some parts of the world, especially in Asia and the Pacific, accounted for the highest rate among the damage resulting from other natural disasters such as landslides, earthquake and tsunamis. Due to this factor, has motivated us to study further on flood forecasting. In previous studies, researchers focus on separate three criteria which are reliability groups, time complexity and error rate to forecast flood. In this paper, we study and analyze the three mentioned criteria in order to identify the highest criteria utilized in flood forecasting. The number of references studied and analyzed is in the year 2010 until 2019. From our findings, the highest criteria identified are under the reliability group, with highest accuracy index of 90%.