<p>In recent years, the issue of drinking water distribution has begun to enter the archives of the population of the city of Bechar, undoubtedly affected by the climate change that has led to the drought of surface water. High dependence on groundwater in arid and semi-arid areas has become a major source, especially drinking. This case enabled us to conduct a study to assess groundwater quality in Bechar using the criteria of 12 physical and chemical parameters. Four predictive models including linear regression, random forests, XGBoost, and AdaBoost were evaluated using performance measures R², MAE, MSE, and RMSE. The results revealed that the measured values of the water quality index fall into the category of good quality, ranging from 54.34 to 79.80, making it potable. EC and TDS were observed to have a strong correlation, while pH, Mg2⁺, and NO₃⁻ were not associated with any other variable. These results showed that linear regression proved high accuracy in predicting minimum, maximum, mean, variance, and standard deviation, as its results were closest to measured values, making it the most reliable. Based on these findings, it is recommended that the linear regression model be used to improve the forecast of groundwater quality in the Bechar region and support informed decision-making in water resources management.</p>