Path analysis is an important study that slices the correlation coefficients between two variables to evaluate whether the relationship between them is of cause and effect. This study aimed to estimate the phenotypic and genotypic correlations between agronomic traits and perform a path analysis in order to identify variables for indirect selection aiming at a higher grain yield. Fourteen soybean F6 lines from the soybean breeding program of FCAV–UNESP, Jaboticabal, São Paulo, Brazil, were analyzed. The experimental design was a randomized block design with three replications. The agronomic traits plant height at maturity (PHM), first pod height (FPH), lodging (Ld), agronomic value (AV), number of pods per plant (NP), number of seeds per plant (NS), and grain yield (GY) were evaluated. Overall, the genotypic correlations were higher than their corresponding phenotypic correlations. The genotypic correlations between grain yield and the traits agronomic value, number of pods per plant, and number of seeds per plant were positive, significant, and of high magnitude. Path analysis showed that the trait number of seeds per plant had the highest direct effect on grain yield, while the trait number of pods per plant had the highest indirect effect through the number of seeds per plant on grain yield