Regression parameter estimation is mainly used to describe the effect of dependent variable to the response variable. On development of it, parameter estimation is also used to case of multivariate regression. Seemingly Unrelated Regression model is one of regression multivariate cases which has especially assumption, i.e., correlation between errors on the multivariate regression. Robust S is one of robust estimation methods. This estimation can be resistant in presence of outlier in the data. In this research, we studied about parameter estimation for Seemingly Unrelated Regression model using robust S. We applied the model obtained to General Electric and Westinghouse data from 1935 to 1954