This work offers a thorough analysis of flood estimating techniques, emphasising current developments and enduring problems in the area. Conventional methods—like statistical approaches and regionalization techniques—are frequently employed, but they frequently run into problems because of the hydrological processes’ assumptions about homogeneity and stationarity. Artificial intelligence and machine learning techniques are two examples of data-driven approaches that have emerged as viable alternatives. Improved flood estimates are the outcome of these techniques, which make use of big datasets and computational algorithms to capture intricate interactions between hydrological, topographic, and meteorological variables. It is also looked into how hydrological models and data from remote sensing may be combined to estimate floods. Technologies for remote sensing, such as radar measurements and satellite photos, offer important data on topographical features, land cover, and rainfall. They improve flood forecasting and flood estimating accuracy when paired with hydrological models. However, issues with temporal variability, data resolution, and model calibration still need to be resolved. The further important component of this research is how flood estimation is affected by climate change. The importance of quantifying uncertainty in flood estimation is stressed. For the purpose of risk assessment, decision-making, and flood control methods, it is imperative to acknowledge the inherent uncertainties in input data, model parameters, and prediction outcomes. This review concludes by outlining the progress and difficulties encountered in flood estimation techniques and highlighting the necessity of multidisciplinary study, better data accessibility, and strong modelling frameworks. Flood risk management in a changing environment will be more successful if the mentioned difficulties are addressed.