The primary goal of this study is to assess and map flood susceptibility in the Balikhlychai watershed, Ardabil province. Specifically, it compares the performance of two machine learning models, Random Forest (RF) and Support Vector Machine (SVM), to identify flood-prone areas and key influencing factors for better urban and infrastructure management. Covering 1,036 square kilometers, the research utilized ten environmental variables, including elevation, slope, land use, geology, distance from river, rainfall, and indices like NDVI, NDWI, and TWI. Satellite imagery from Landsat 8 and Sentinel was processed in ArcGIS 10.8. The models were implemented to predict susceptibility, with variable importance measured via Gini impurity reduction for RF and variable displacement for SVM. Model accuracy was evaluated using the ROC curve (AUC index). Findings revealed that the RF model (AUC = 0.8) significantly outperformed the SVM model (AUC ≈ 0.6). Decreased vegetation density correlated with a 30-35% rise in flood potential. While "slope" and "elevation" were identified as the most critical factors in the RF and SVM models respectively, high-risk zones were concentrated in the northeastern Ardabil plain due to topography and human activity. The study confirms that morphometric characteristics and vegetation cover heavily influence hydrological behavior.