Abstarct: Whenever there are no other resources to provide water, groundwater resources are the only source in many regions. Reduction in groundwater, pollution and improper management are some issues that affected human life. Considering the role of water in human health, its quality must be inspected via conducting researches for quality control. In this area, smart and data-centered models are rapidly developing methods due to their ease of use compared with others. These models can learn and generalize data and are able to solve problems of prediction, management, and estimation in various water management aspects. Meanwhile, neural networks are powerful tools in solving regression and classification problems that can solve problems with complex hydrology and low data compared to data-centered methods. In this study. a feedforward neural network method and support vector machine (SVM) method is employed to predict temporary and permanent hardness concentrations of Amol-Babol plain using chemical parameters of water, such as , , , , , , and . For this purpose, the qualitative data collected from 855 wells during two fall and spring seasons in six years was used. Then, by creating a feedforward neural network method and SVM model with different parameters, the optimal model was selected, trained, and tested. baxsed on R2, MSE and RMSE criteria, the obtained results were 0.9303, 748.76, and 27.36, respectively, for feedforward neural network at the permanent hardness and 0.9688, 504.78, and 22.46, respectively, for temporary hardness. These results are, respectively, 0.9304, 0.0001401, and 0.0181 for permanent hardness and 0.9833, 0.0000338 and 0.0058, respectively, for temporary hardness in the SVM. The obtained results of temporary and permanent hardness in Amol-Babol plain by data-centred models showed the capability of created models in modelling, which predicted hardness concentration in the desired area with an acceptable approximation.