Evaluation of groundwater quality remains essential for irrigation; however, in agricultural countries, financial limitations often lead to inadequate sampling frequency, which hampers thorough evaluations. This study aims to assess the irrigation water quality index (IWQI) in the Northern Gab & egrave;s Aquifer, Tunisia using two usual methods and different models supported by machine learning techniques. Hydrochemical parameters and geospatial analysis were conducted to assess the groundwater suitability for irrigation in the study area. Additionally, Machine Learning models, including Extreme Gradient Boosting (EGB), Support Vector Machines (SVM), and Resource Description Framework (RDF) were applied to predict IWQI. 42 groundwater samples were analyzed to calculate the IWQI. According to the first traditional method, the findings revealed that 35.714% were deemed unsuitable for irrigation. Whereas the second method shows that 92.857% of the samples were considered medium suitability for irrigation. Conversely, obtained results show that SVM performs exceptionally well for the two methods in terms of precision and consistency, achieving a high R2 values exceeding 0.99, and low RMSE and MAE. It demonstrates impressive predictive ability with strong performance. These differences in performance between these machine learning models bring great implications for making decisions in groundwater management in the study region. Implementing sustainable water management methods, such as improving irrigation and encouraging alternative water sources, is critical for the enduring accessibility. Future research should focus on integrating real-time monitoring systems, advanced machine learning algorithms, and geospatial technologies to enhance the accuracy and efficiency of groundwater quality assessment for sustainable irrigation practices.Graphical AbstractThis is a visual summary serves as a pivotal entry point into the research, offering a concise overview of the study's core findings and methodologies. It illustrates how AI models contribute to sustainable water resources management. It depicts the process from water sampling and chemical analyses, through both traditional and AI-driven modeling, to the final evaluation using the Improved Water Quality Index (IWQI). It begins with "Water Sampling and Chemical Analyses", indicating the experimental collection of data. These data are then used in "Traditional Models", suggesting a comparative approach. The ultimate output is the "IWQI Assessment," which stands for Improved Water Quality Index, calculated using two different methods with predefined formulas and weighted parameters to assess the overall groundwater quality. These "Chemical Parameters" are also fed into "AI models" for "Prediction and Classification". The Artificial Intelligence (AI) models process the data ("Data processing") using "Machine learning" techniques (ML) such as Extreme Gradient Boosting (EGB), Support Vector Machines (SVM), and Resource Description Framework (RDF). The results are then validated through machine learning "Model Validation". The significance of this process is highlighted in the two graphs comparing "Traditional Models" with "Simulated IWQI (SVM)", since the SVM model is validated by its superior performance over EGB and RDF models. Overall, this graphical abstract summarizes the paper's methodology, data processing, and comparative analysis, providing a snapshot of the research's scope and findings. This AI-driven assessment is contrasted with traditional modeling approaches, highlighting the enhanced accuracy and efficiency that AI offers to predict water quality and emphasizes the power of the AI models and their capacity to support sustainable water management.