The aim of this research is to optimize the stable portfolio based on the weighted STARR risk-reward ratio model. Uncertainty in stock returns is modeled by changing them in symmetric bounded intervals. The robust reward-to-risk ratio used preserves linearity in the resulting models, and hence are solvable with acceptable efficiency. In this research, the price data of ۲۰ active stocks in the Tehran Stock Exchange during the years ۲۰۱۴ to ۲۰۲۱ have been used. Using these data, ۴۰ portfolio optimization problems (۲۰ normal problems and ۲۰ robust problems) have been solved using GAMS software and CPLEX solver. Based on these data, the performance of robust and normal STARR models are compared based on the criteria of value at risk (VaR), conditional value at risk (CVaR), Sharpe ratio, standard deviation, and returns. The results show that the portfolios formed using the robust model have better risk measures compared to the normal model.