This research is aimed at determining and introducing a suitable model for investment decision in security market. To this end, effective measures are extracted for portfolio selection according to research literature. Therefore, the numbers of basic variables are determined with regard to the fact that there is correlation between financial ratios. In the following data envelopment analysis is used to rate understudy population including fundamental metals, metallic minerals, chemicals, medicine, properties and real states and cement industries. For the reason, 3-year average of real data in the period of 2006-2008 is used. After rating company of each mentioned industry, those companies having efficiency above 0.9 are located in the selected portfolio. 3-year output of portfolio of 2006-2008 is calculated to optimize selected portfolio and then it is compared with portfolio of unselected companies. Finally, genetic algorithm is used to evaluate optimization of selected portfolio according to Terner scale using optimization of the mentioned portfolio. Results show that the portfolio selected from the companies having higher efficiency leads to better output compared with unselected companies. This model is fitted to designate investors’ resource and gain more output.