Today, the flood phenomenon is one of the most complex hazardous events that, more than any other natural disaster, causes human and financial losses and agricultural land destruction every year in different parts of the world; therefore, preparing a flood sensitivity map is the first step in a flood management program. The aim of this study was to identify flood-sensitive areas using two machine learning models: Random Forest (RF) and Support Vector Machine (SVM) in the Navrud watershed in Gilan province. A map of the distribution of past floods was prepared in order to predict future floods. Out of 683 flood events, 70% (466 flood events) were used for modeling and 30% (200 flood events) were used for validation. By reviewing previous studies and surveying the study area, 10 effective factors were selected and prepared for flood zoning, including distance from the watercourse, drainage density, flow direction, slope, slope direction, precipitation, elevation, land use, and geology. The results of the sensitivity analysis showed that the three factors of drainage density, geology and land use have the greatest impact on the flooding of the study area, respectively. Also, the results of the model output evaluation showed that the AUC value in the SVM and RF models was 0.97 and 0.93, respectively, which indicates the superiority of the RF model and its greater accuracy in preparing a flood sensitivity map in the study area. The largest flood sensitivity area in the RF model is related to the low class and in the SVM model to the middle class.