The treatability of high sulfate water using elecrocoagulation method was modeled in present research. Electrocoagulation unit with continuous flow reactor was constructed. Synthetic samples of different sulfate concentrations (1000, 1750, 2500, 3250) mg/l were used at current (0.5, 1.0, 1.5 and 2.0) A, reactor detention time (37.5, 75, 112.5, and150) min, and effective surface area of aluminum anodes (0.0096 and 0.0324) m2. Statistical models were developed to document and evaluate the relationship among the independent variables and the dependents variables. In addition, optimization technique was applied on the necessary statistical models to find the optimum current and time required to achieve a certain removal percent at minimum operating cost. Furthermore, the most applicable kinetic and isotherm models were examined for the experimental data to determine the most suitable representative models. The multiple linear regression relieved that all the models had very high significance at level of confidence (1.0) %. The optimization showed that the minimum operating cost occurred at optimum current of (0.5) A and the optimum time increased as the removal percent increase until reached upper limit value of (150) min then the optimum current begin increase. The first and second order models were proper to represent the kinetic adsorption of magnesium sulfate by aluminum hydroxide flocs at steady state condition. The Freundlich isotherm model was the most model matching the experimental data at applied current of (0.5)A while the Langmuir model was the most model matching the experimental data at applied current greater then