The dissolved oxygen concentration (DO) is of utmost importance in maintaining a healthy aquatic ecosystem. However, escalating industrialization and population growth have led to a decline in DO concentration, causing contamination of India's rivers, most notably the Godavari River. The Godavari River basin is experiencing this issue due to the introduction of new industries, unsustainable farming practices, and religious sites. This study sought to monitor DO water quality downstream (d/s) of the Godavari River in Rajahmundry (Dowlaiswaram), East Godavari district of Andhra Pradesh. Using soft-computing techniques, specifically artificial neural network (ANN) and random forest (RF), we developed models to predict DO using other critical water quality parameters, including temperature (T), pH, and electrical conductivity (EC). Four input water quality datasets, namely temperature, pH, EC, and DO, were employed to construct the predictive models. Statistical metrics such as regression coefficient (R2), root mean square error (RMSE), and mean arithmetic error (MAE) were used to evaluate the predictive efficacy of ANN and RF models for different combinations of input data. The paper concludes with an analysis of the results obtained from the ANN and RF models and insights derived from them.