Agriculture, a fundamental pillar of national economies, is highly susceptible to environmental factors such as climate and weather conditions. Crop productivity efficiency relies on various agricultural variables, water availability, temperature fluctuations, and climate variations. Addressing the difficulties in agriculture, particularly in prediction and water quality assessment, requires innovative technological solutions. In order to improve production targets and support sustainable agricultural practices, this research compares and contrasts machine learning and gradient boosting methods for forecasting the water quality of potato crops. The proposed system centers on the analysis of water quality parameters crucial for optimal potato growth. Leveraging a comprehensive dataset containing metrics like pH levels, solids, organic carbon, and sulfate, the system provides valuable insights into the specific water quality requirements necessary for cultivating healthy and high-yielding potato crops. The integrated machine learning techniques include k Nearest Neighbor (kNN), Logistic Regression (LR), Random Forest (RF), Decision Tree Classifier (DTC), Gaussian Naive Baye’s (NB), and Support Vector Classifier (SVC). Additionally, gradient boosting approaches such as CatBoost Classifier (CBC), XGBoost (XGB), and LightGBM (LGBM) play pivotal roles in the prediction process. The study’s findings indicate that the CatBoost Classifier yields higher accuracy compared to other approaches. This research contributes valuable knowledge that could inform stakeholders, including researchers, farmers, and policymakers, enabling evidence-based decisions that promote sustainable and efficient potato cultivation. This research enhances the comprehension of precision agriculture by clarifying the connections between water quality factors and the results of potato cultivation, enabling well-informed approaches to improve resource use and boost crop yield sustainably.