This paper evaluates the contribution of human capital for Brazil's agricultural growth during the time period from 1970 to 1995/96. Two different versions of neoclassic models are used: the traditional model and the alternative one. The econometric results indicated that human capital's effect on the agricultural output growth was positive and highly significant in all equations estimated. By separating the human capital data set with the purpose to evaluate the existence or absence of threshold effect, it was verified that evidence does not exist in order to support the existence of that effect into the Brazilian agriculture. The catch-up variable in the alternative neoclassical model was not significant and its coefficient was near to zero. This result does not support the existence of technology diffusion among the Brazilian states' agriculture.