In this study, three intelligent modeling approaches, including Adaptive Neuro-Fuzzy Inference System, Artificial Neural Network–Multilayer Perceptron, and Artificial Neural Network–Radial Basis Function, were employed to simulate 14 water quality parameters, in the Halil River over a long-term monthly dataset spanning 52 years. Streamflow discharge was used as an input variable in some models. The performance of the developed models was evaluated using statistical indicators including the coefficient of determination (R²), root mean square error (RMSE), mean absolute error (MAE), and Willmott’s index of agreement (d). The results indicated that although all three artificial intelligence models demonstrated high capability in simulating the investigated water quality parameters, the ANFIS model outperformed the other models in most cases. Specifically, for SAR, values of R² = 0.99 and RMSE = 1.04 were obtained; for %Na, R² = 0.99 and RMSE = 3.02; and for SO₄, R² = 0.93 and RMSE = 0.61. Among the input scenarios, model (b), which incorporates five easily measurable parameters, is recommended as a practical and efficient model, as it reduced RMSE by approximately 65% compared to model (a) (based solely on discharge). The findings of this study demonstrate that intelligent machine learning techniques can be effectively used to estimate water quality parameters in cases where direct measurements are unavailable, based on other measured variables, thereby reducing laboratory analysis costs.