Application of selection index for screening desired plants based on complex quantitative traits can be more effective than direct selection. This study was conducted using an F₂ rice population consisting of 87 plants derived from a cross between two cultivars Gharib and IR28. The purpose was to establish suitable selection indices for increasing yield and its related traits in research farm of Rice Research Institute of Iran (RRII), Rasht, in 2005. Studied traits included the days from sowing, germinated grain to maturity (MD), plant height (PH), panicle length (PL), flag leaf length (FL), flag leaf width (FW), number of panicles per plant (PP), number of grains per panicle (GP), number of spikelets per panicle (SP), 100-grain weight (GW), grain yield per plant (GY), biomass (BM), harvest index (HI), grain length (GL) and grain breadth (GB). Among the studied traits, 100-grain weight (GW), biomass (BM) and harvest index (HI) (0.99) and flag leaf width (FW) (0.35) showed the highest and lowest broad-sense heritability, respectively. Path coefficient analysis revealed that BM, HI, GP had positive direct effects on GY. Calculation of five different selection indices based on optimum and base indices indicated that selection for BM, HI and GP using genotypic path coefficients and their heritability as economic values would be a suitable selection criterion for improving population. Moreover, this study showed that both optimum and base indices show the same genetic progress for the studied traits. Since evaluation of base index, is much easier than the optimum index, it is highly recommended.