The present study aims to estimate the flood susceptibility degree over the Prahova River basin located in the central-southern part of Romania. To obtain the proposed outcomes, the next 10 flood predictors were used as independent variable in the machine learning models: slope angle, convergence index, distance from river, elevation, plan curvature, hydrological soil group, lithology, topographic wetness index, rainfall and land use. The factors along with 158 flood locations, that represent the dependent variables, were involved in the training procedure of the following four ensemble models: Deep Learning Neural Network –Statistical Index (DLNN-SI), Particle Swarm Optimization-Deep Learning Neural Network–Statistical Index (PSO-DLNN-SI), Support Vector Machine–Statistical Index (SVM-SI) and Particle Swarm Optimization-Support Vector Machine–Statistical Index (PSO-SVM-SI). Through the Statistical Index method, the coefficients of each flood predictor class/category were calculated. The best performance was achieved by PSO-DLNN-SI model for which an AUC-ROC Curve of 0.952 was calculated. It is worth to note the application of PSO algorithm manage to increase the models performance. Also, it is important to note that around 25% of the study area has a high and very high exposure to flood phenomena.