The objectives of this study were to obtain information about genetic parameters of a partial diallel in common bean, as well as to evaluate the effects of growing seasons, locations and generations in these estimates in order to define the best strategies for the use of diallel analysis as methodology of selection of parents and populations in common bean breeding. Twelve parents were crossed in a partial diallel scheme (5 × 7), and the F1 and F2 generations were evaluated in four experiments (I, II, III, and IV). The effects of seasons were determined in the Experiments I and II, in which the F1 generation was evaluated in two distinct growing seasons. To quantify the effect of locations, we used Experiments III and IV, in which the F2 generation was evaluated in the same season, but at different locations, and to quantify the effect of generations, the experiments II (generation F1) and III (F2 generation) were evaluated in the same season and location. As for grain yield the genetic control varied according to the environment. An effect of interaction on the genetic parameters of the diallel was observed for generation, whereas for grain yield, effects of season and location were observed. For plant architecture, parental combining ability should therefore be estimated using the F1 generation, while for grain yield, both the F1 and F2 can be used. The recombination between the best hybrids is a promising strategy for extraction of superior black bean lines.