One of the main aims of water resource planners and managers is to estimate and predict the parameters of groundwater quality so that they can make managerial decisions. In this regard, there have many models developed, proposing better management in order to maintain water quality. Most of these models require input parameters that are either hardly available or time-consuming and expensive to measure. Among them, the Artificial Neural Network (ann) Models, inspired from human brain, are a better choice. The present study has simulated the groundwater quality parameters of Ramhormoz Plain, including Sodium Adsorption Ratio (SAR), Electrical Conductivity (EC), and Total Dissolved Solids (TDS), via ann and ann+ Particle Swarm Optimization (PSO) Models and at the end has compared their results with the measured data. The input data for TDS quality parameter is consisted of EC, SAR, pH, SO4, Ca, Mg, and Na, while for SAR, it includes TDS, pH, Na, and Hco3, and as for EC, it involves So4, Ca, Mg, SAR, and pH; all of them, gathered from 2009 to 2015. Results indicate that the highest prediction accuracy for SAR, EC, and TDS is related to the ann + PSO model with the tangent sigmoid activation function so that both MAE and RMSE statistics have the minimum and R2 the maximum value for the model. Also the highest prediction accuracy is respectively related to EC, TDS, and SAR parameters. Considering the high efficiency of artificial neural network model, by training the PSO algorithm, it can be used in order to make managerial decisions and ensure monitoring and cost reduction results.