TheMonitoring MahanadiGeographical Information System River, OdishaMahanadi River, Odisha, India, possesses strong links to the locals both traditionally and economically. It offers enough arable plains and hills for farming, as well as suitable areas for communal hunting and fishing. Because of the abundant natural resources it offers, it is significant not only for the residents of particular areas but also for the state of Odisha. However, increased organic matter content downstream of watersheds highlighted the seriousness of non-point source (NPS) pollution. So, the assessment of the pertinent physicochemical parameters of 19 locations of surface waterSurface water quality from the river basin is taken into account, for a period of 8 years (2015–2023). The standard guideline values were compared to the physicochemical results as advised by the World Health Organization (WHO) regarding water consumption and public health, to have a general overview of surface waterSurface water quality in this recent scenario. The results further indicated that the majority of the parameter values fall below the highest amounts that are allowed to be consumed in accordance with the WHO guidelines except TKN and TC, during summer season. Using geographical information system (GIS)—Inverted Distance Weighting (IDW) methods with ArcGIS 10.8 application software, spatialSpatial distribution maps have been created. To identify and allocate the sources of pollution and analyse the spatiotemporal variation, the dataset was exposed to a widely accepted coherent approach employing Level Based Weight Assessment (LBWA/L) Water QualityWater quality Index (WQI) and Multi-Criteria Decision-MakingMulti-criteria decision-making (MCDM) analysis, coupled with Machine Learning (ML) techniques, namely Decision Tree (DT) Algorithm. It is observed that samples mostly contained the anions \({\text{Cl}}^{ - } \, > \,{\text{SO}}_{4}^{2 - } \, > \,{\text{NO}}_{3}^{ - } \, > \,{\text{F}}^{ - }\) , while the primary cations were Fe2+ > B+. For each and every sampling site, L-WQI was established to be 23–412, respectively. It is found that most of the test sites were located in excellent to unsuitable category, as determined by this approach. The results revealed that 36.84% samples are excellent for drinking, 10.53% good, 31.58% poor, 15.79% very poor, and 5.26% unacceptable and, thus, unsuitable for consumption. Further, at sites X-(2), (5), (6), (7), (8), (9), (10), (11), (12), (13), and (19), water qualityWater quality is mostly impacted by agricultureAgriculture and industrial inputs. In addition, sample sites situated in the upstream and downstream regions see declining water qualityWater quality indicators as a result of hydroelectric dams, evolving land use patterns, growing populations, and deforestation along riverbanks and catchment areas. The present study points out from the spatialSpatial distribution maps that \({\text{EC}},{\text{ Cl}}^{ - } ,{\text{ SO}}_{4}^{2 - } ,{\text{ NO}}_{{3}}^{ - }\) , TKN, and TC played a major part in influencing the river's WQI, showing that during the monsoon season, both insignificant human actions and natural causes have an effect on surface waterSurface water chemistry. However, to find a better alternative in estimating the level of rankingRanking of pollution, a mathematical technique DT has been adopted. This will offer helpful information to prioritise and preserve the water qualityWater quality of drinkable sources and lessen the negative effects that consuming subpar surface waterSurface water resources have on human health. It is also adaptable, impartial, simple to compute, and time-saving. Moreover, the results showed that the score and its rank at X-(9) (0.95), signified as mostly polluted site, in comparison to other locations. The cause may include human activities, prolonged wastewater use, excessive surface waterSurface water extraction, and modifications to land use patterns. Ultimately, both LBWA and DT are permitted to identify elemental correlations, and their probable origin was man-made as opposed to nature. These findings are crucial to comprehending the study area's surface waterSurface water sustainability for drinking. Hence, the present study will serve as the foundational database for all upcoming work in the area. The recommended mitigation measures include debris removal, silt extraction, river bank stabilization, modern hydraulic structures, improved waste management, systematic removal of water hyacinth and decomposed materials, and spoil bank design in spilling zones to restore the river's natural flow.