Groundwater is an important source of freshwater for drinking, irrigation, and industrial purposes in many regions. However, groundwater systems are increasingly threatened by quality deterioration driven by intensive agriculture, urbanization, and land use changes. Nitrate and chloride are common contaminants which are key indicators of anthropogenic pressure on aquifer systems. Elevated nitrate concentrations, largely originating from fertilizer application and manure management, pose serious risks to human health and aquatic ecosystems. Also, high chloride levels, often associated with road de-icing salts, can deteriorate water quality and alter subsurface geochemical processes. Machine learning (ML) techniques offer a powerful data-driven alternative for accurate prediction of these ions in groundwater systems by capturing the nonlinear relationships between the involved parameters such as hydrogeology, climate, and land-use characteristics. In the current study, different regression-based ML algorithms (e.g., support vector regression) were adopted to predict nitrate and chloride concentrations in groundwater in different hydrogeological and land use settings. The optimal ML models were selected based on a group of evaluation metrics such as root-mean squared error. In addition, the interdependence between the involved process parameters (e.g., hydrogeological conditions) and groundwater quality was interpreted to determine the governing parameters on the contaminant transport process in sub-surface water. The main outcomes of this study can help decision-makers in developing the most effective groundwater management strategies to protect and improve groundwater quality. In addition, these insights enable the interpolation of nitrate and chloride concentrations from discrete sampling points, facilitating predictions at unmonitored locations across the watersheds.