Grain yield is a quantitative trait advance in which through direct selection would be time consuming. Indirect selection employing selection index is more promising.A field experiment was conducted during the 1999-2000 growing season in two locations (Kooshkak and Badjgah). Thirteen corn hybrids were employed in a randomized complete block design with three replications in each location. Thirty five traits were assessed five developmental stages (stem elongation, tasseling, blister, hard dough and physiological maturity), combined analysis of variance and covariance being conducted. Finally 12 traits (grain yield, CGR4, Ear length, NAR2, NAR3, LAI2, LAI4, Row number per ear, Internode length in 3rd stage, Plant height in 3rd stage and coated ear weight in 5rd stage) were selected for constructing selection indices via path analysis. First selection index was conducted in Robinson,s method. Selection index including grain yield, NAR2, LAI2, row number per ear and node number in 3rd stage was found to be the best index. Genetic gain index was the 2nd kind used in this investigation. Selection index including LAI4 and node number in 3rd stage was evaluated as the best index. In both selection indices, physiological indices including net assimilation rate, crop growth and relative growth rate were found to be the most important traits among the best selection indices.