Surface water is an essential water source in arid and semi-arid regions. Human activity is having a growing effect on surface water. After many years, the investigation must address two crucial questions: the water quality and its formation mechanism. The Mahanadi River, Odisha, is surrounded by 55% agricultural land, and farmers only get rainfall in this location. The current analysis is to evaluate the suitability of river water for different purposes, such as drinking and agriculture. The present study is investigated by employing long-term water quality monitoring data, for a duration of 2021–2023, obtained from 19 selected sites. The extensive, site-specific analysis provides a distinctive perspective through which the environmental issues at each site are analysed. It is further enhanced by the wide range of community inputs. The study’s results showed that most of the parameter values are under the maximum permissible limits as per WHO except TKN Fe2+ and TC. This study also demonstrated how surface water quality may be assessed for drinking purposes by utilizing entropy-based WQI, with reliability-based machine learning (ML) models such as ensemble ranking (ER) have been employed. Hence, these two models act as a benchmark model for predicting WQ values at the watershed. Finally, the analytical findings were used to produce the parameters’ numerical geographic distribution using the geographical information system (GIS) environment. The E-WQI revealed that around 37% of investigated sites fall under poor/very poor water quality, while 52.63% of obtained locations indicated under excellent water class. However, except for seven samples, all other samples were categorized as excellent-medium for human consumption. Entropy’s findings showed that the main causes of fluctuation in water quality were fertilizer, organic waste, and soil leaching. The findings once more demonstrated that the river water has highly acceptable water quality features in many places and that the heavy elements, specifically Fe2+, found in the water did not exceed the thresholds, which frequently made them suitable for home and agricultural use. As a result, to find a better alternative in estimating, ER was implemented to reduce inconsistencies that involve the WQI index. The proposed approaches depicted U-(9), was termed as the most polluted site, when contrasted with other places. In this context, the other affected places in the current study are U-(8) and U-(19). The degraded water quality at the bad sites was found to be caused by the growing and varied types of human activities, as well as the discharge of industrial and agricultural waste along the river flow. The results of the ER indicated that the degradation of the water quality in the upstream and downstream zones is caused by residential wastewater and leachate contamination at sampling points U-(9), (2), (19), and (8). However, water needs to be treated prior to consumption at the alarming stations that are prone to be polluted. The outcomes of this investigation have the potential to enhance comprehension of the mechanisms governing the chemistry of surface water and serve as a benchmark for comparable areas across the globe. At the end, a comprehensive understanding of water quality issues is made possible by combining scientific data with local perspectives. This highlights the critical need for focused interventions to protect the river ecosystem and the welfare of the communities that depend on it, and it also offers insightful information for sustainable environmental management.