Abstarct: Surface water identification and monitoring and flood tracking using remote sensing science have taken much attention in recent decades due to their importance in satisfying human needs and political decisions. Therefore, surface water and flood tracking have been done in this study using remote sensing systems and Sentinel-1, Sentinel-2, and Landsat-8 satellites. In this thesis, two scenarios were used to identify surface waters and then flood with high accuracy. Both scenarios consist of two approaches of data fusion and decision fusion to increase the accuracy of surface water identification using the fusion of results. Therefore, after increasing the spatial resolution of Landsat-8 images from 30 meters to 10 meters in the first scenario, water indices such as NDWI, MNDWI, AWEI, and WI2015 were extracted. Then by combining the information obtained from the sensors, the multispectral bands with Sentinel-1 data are stacked in one laxyer. The output results of classifiers such as SVM, NN, and RF, representing the data fusion approach, are fused by the decision fusion approach and the maximum voting algorithm. In the second scenario, using the approach of the first scenario, surface water is identified by subtracting the results from before and after the flooded data, the stagnant water is removed, and the floodwater is accurately identified. Finally, the results achieved from the SVM, NN, RF, and KNN classifiers are combined using the Maximum Voting Algorithm and the Dempster-Shafer algorithm. The accuracy of the output results is considerably improved. The accuracy obtained from both scenarios shows that the decision fusion approach has increased the accuracy of the final results. In the first scenario, we see that the decision fusion approach has increased the overall accuracy by 2% compared to the RF classifier, which had better accuracy than the other classifiers, from 96.87 to 98.86% for the Gelvard Dam and from 96.73% to 98.56% for the city of Neka to identify surface water. Also, in the second scenario, the decision fusion approach showed an accuracy of about 1.5% higher than the data fusion approach. This approach increased the overall accuracy of the RF classifier for the Sentinel-1 and Sentinel-2 flood images from 91.51 and 94.61% to 94.89% for Majority Voting and 95.71% for the Dempster-Shafer, respectively. The two proposed scenarios clearly show that the decision fusion approach for both study areas has increased the accuracy of surface water identification and flood track.