We developed a flood simulation model capable of evaluating the flood control function of paddy field dams and proposed a method for mathematically exploring the optimal placement of these dams using mathematical programming. The flood simulation model employs an unstructured grid flood simulation model with added functionality for paddy field dams. Genetic algorithms were used to search for the optimal placement of the dams, with their on(1)/off(0) status represented as a gene sequence in the algorithm. The model was applied to a specific area in the Takatoki River basin, which includes an open levee and multiple paddy fields at the confluence of rivers. The results demonstrate that the genetic algorithm enables an efficient search for configurations that reduce water levels compared to exhaustive search approaches. For rainfall events that do not exceed the storage capacity of the paddy fields, the reduction in water levels due to paddy field dam placement is low, suggesting that the effects can be expected regardless of the placement. On the other hand, for rainfall events that exceed the storage capacity of the paddy fields, the reduction in water levels due to paddy field dam placement becomes significant, indicating the importance of considering the placement of the paddy field dams.