The selection of genotypes of soybean with high grain yield and high capacity of environment adaptation are the objectives of the genetic improvement programs. To verify if the genetic gain through of the breeding of the soybean crop is being reached it is necessary that, frequently, each improvement program makes an auto-evaluation to verify if there has been effective progress, if a certain culture condition was privileged by the breeding and to correct eventual distortions of goals. To have significant advances the breeding programs have to evaluate their efficiency and it is necessary to use methods that quantify the genetic advances. The possibility of prediction of the gains obtained by selection consists in one of the main contributions for the breeding, becoming possible to guide in a more effective manner the breeding program, to predict the success of the adopted selection method and to determine, scientifically, which techniques can be more efficient.