Discharge has traditionally seen direct applications of itself for water management. From river embankments to dams, to design and model all, water discharge is required for long-term adaptation to climate change’s effects, including floods. Floods are one of the most common and economically costly natural disasters that, because of climate change, have only intensified in recent months. Predicting floods requires a multitude of factors and parameters, each having its own complexity to deal with, among which, river discharge is always used as a parameter. The advent of technology has led to the creation of many predicting models, but because of the continued advancement of artificial intelligence, new techniques are tested to find more accurate results. Although there are statistical models using which floods are predicted, models that utilize machine learning are increasingly being adopted as Academia is progressively moving towards ML techniques. This work includes numerous Machine Learning algorithms that have been applied to predict river discharge on a time-series dataset. We ran our experiments on the publicly available atmospheric, land, and oceanic climate variable database called ECMWF Reanalysis v5 (ERA5). The river discharge variable was obtained from GloFAS; both are made available by Europe's Copernicus Climate Data Store (CDS). We processed the data over time and the shape of the river basin to allow predictions at a more precise scale. We show that Machine Learning techniques are a good choice in terms of model complexity and performance on this dataset. The top three performing machine learning models are Support Vector Regression, Gradient Boost, and Random Forest. But the only neural network model that is used, Time Delay Neural Network outperforms all the other models.