The fresh water resources are limited in the world and their distribution is heterogeneous. In the other hand industrial establishment and dam construction decrease the quality of these limited resources, so the monitoring of water quality is necessary and unavoidable. There are different methods for consideration of water quality changes ) statistic tests, regression, time series, multivariate methods…). In this study, monthly time series of water quality parameters (SO 4 -2 , Cl - , HCO 3 - , pH, EC, SAR) are used for study of Zayandehrud water quality. For this purpose, procedure presented by Box Jenkins were used for modeling which contain three stages, including model recognization, parameter estimation and control of estimated parameters. For all parameters in six hydrometric stations, ARIMA, ARMA , AR , MA, multiplicative or seasonal models were obtained. However, SO 4 -2 in Ghaleh Shahrokh station, HCO 3 - in Eskandari station, EC in Polekale station, pH in Polechoom station, SAR and EC in Varzaneh station were not modelled. After modeling pH, Cl - and HCO 3 - were forecasted from shahrivar 1384 up to shahrivar 1386 and the percentage of error was calculated. The lowest percentage of error was obtained for pH parameter. In additional to modeling, the correlation between the river discharge and other parameters was calculated and it was distinguished that discharge has negative correlation with all parameters except pH in Eskandari station.The mean of discharge in twenty years period showed that Ghaleh Shahrokh station has the highest and Varzaneh station has the lowest mean of discharge. Key Words : water quality, modeling, ARMA and ARIMA models, time series, Zayandehrud