<p>Flood hazard maps (FHMs) are produced using various methods. In semi-quantitative approaches, the accuracy and reliability of these maps depend on data quality and the models employed. This study assesses the effectiveness of the valley floor width index to improve the accuracy of flood hazard maps. Also, we compare different multi-criteria decision-making (MCDM) weighting methods including Shannon entropy based on probability density (ED), Shannon entropy based on original data (EO), the Best Worst Model (BWM), and the Method based on the Removal Effects of Criteria (MEREC) for flood hazard mapping in Tajen basin (Iran). Factors such as elevation, slope, ruggedness index, distance from the river, rainfall intensity, lithology, and land use were used for flood hazard mapping. These factors were rated using expert knowledge and fuzzification. In order to limit the distance buffers based on topographic conditions, we combined factor of distance from the river with the valley floor width index. A flood inventory map containing 170 flood points was used with a ratio of 70–30% for training and testing, respectively. Results demonstrated that the ED weighting method has the highest accuracy compared to other methods. Additionally, the fuzzy rating method indicated higher accuracy compared to the expert knowledge method. We conclude that employing the valley floor width index increases the accuracy of flood hazard mapping in the study area.</p>