Density is an important indicator of vegetation evaluation, which several methods have been developed for its assessment, but their accuracy is concerned. To reveal accuracy of each methods, a study site of 32000-m2 in the steppe rangeland of Marjan, Boroujen was selected and divided into eight macroplots of 4000-m2. Then, the individuals of Astragalus verus and Astragalus albispinus were counted in each microplots. A transects of 100-m established parallel to the length of the macroplots (40 × 100 mm) and density was measured using closest individual, nearest neighbor, random pairs, point-centered quarter, third closest individual, angle order and variable area transect methods, systematic-randomly in 10 sampling points in each macroplots. The results indicate that the real density of A. verus and A. albispinus were 0.1593±0.084 and 0.0622±0.0282/m2, respectively. Closest individual (0.1357±0.1315/m2), nearest neighbor (0.1368±0.1432/m2), point-centered quarter (0.1016±0.1664/m2), random pairs (0.0588±0.0536/m2), third closest individual (0.1107±0.0775/m2) and variable area transect (0.0221±0.0105/m2) for A. verus and angle order (0.0927±0.0523/m2), nearest neighbor (0.0424±0.0357/m2) and third closest individual (0.0524±0.0447/m2) for A. albispinus showed no significant difference with controls. The results revealed that the closest estimation to the controls belongs to the nearest neighbor (-0.141) and third closest individual (-0.0098) for A. verus and A. albispinus, respectively. Moreover, the nearest neighbor (RMSE=0.6877, SE=0.0026 and R=0.1147) and closest individual (RMSE=0.5609, SE=0.0007, R=0.0320) showed the most precise estimation of densities of A. verus and A. albispinus, respectively. Generally, the closest individual for estimating plant density of A. verus and the nearest neighbor's for A. albispinus are proposed.