Water is essential for the survival of humans, animals, and plants, yet its quality often falls short of suitability for various purposes due to industrialization, mining, pollution, and natural factors altering its composition. Regulatory bodies like the World Health Organization set guidelines for acceptable water quality levels, emphasizing parameters crucial for human consumption and irrigation. Assessing water quality involves laborious sampling, parameter measurement, and adherence to stringent guidelines, presenting challenges. This study proposes a network architecture utilizing LoRa technology to collect real-time water parameter data, aiming to automate the assessment of water suitability for drinking and irrigation using machine learning (ML) tools. Simulations employing Radio Mobile suggested a partial mesh network topology as most effective for the monitoring network, considering land topology. Given the scarcity of large, open datasets for drinking and irrigation water, the study developed usable datasets for ML model training. Three ML models— Random Forest (RF), Logistic Regression (LR), and Support Vector Machine (SVM)—were evaluated for water classification, with LR showing superiority for drinking water and SVM for irrigation. Recursive feature elimination was employed with the ML models to identify the most influential water parameters on classification accuracies