We estimated the genetic distances among 10 spring wheat genotypes based on pedigree data, morphological traits and AFLP markers, used individually and combined with morphological traits, to find the best predictors of general- and specific-combining abilities among parental genotypes. Ten wheat parents were crossed in a diallel form, disregarding reciprocal hybrids, totaling 45 combinations. The F₁ hybrids, F₂ populations and parents were evaluated in the field in 2007. The experimental plots consisted of 20 plants for F₁ hybrids and 40 plants for parental and F₂ populations. All methods (pedigree data, AFLP markers and morphological traits, used individually and combined) were found to be useful for the assessment of genetic diversity. The significant coefficient correlations ranged from low (0.45) to moderate (0.67) between the distance measures and hybrid performance. There was significant agreement between the distance measures based on AFLP markers vs morphological traits + AFLP markers (r = 0.47) and between pedigree data vs morphological traits + AFLP markers (r = 0.43). The pedigree distance was positively associated with traits 100-kernel weight and grain yield per plant in F₁ (correlations of 0.67 and 0.62, respectively) and F₂ (correlations of 0.62 and 0.59, respectively) generations. These correlation values indicate that the genetic distance, based on pedigree data, could replace diallel crosses for the selection of parents with higher combining ability and with moderate reliability.