Abstract: In 2011, a large-scale (72 ha) and long-term experiment was established in Sinop, Mato Grosso, Brazil, with the objective of evaluating different grain production systems. All systems with the presence of trees were implanted in the configuration of triple groves eucalyptus with spacing of 3.5 m x 3.0 m x 30 m (270 trees/ha). After the 5th year, we made 2 types of thinning: selective thinning, removing 50% of the trees, maintaining the configuration of triple lines (135 trees/ha); systematic thinning, with removal of the two side lines (90 trees/ha remaining). After the 8th year, the systems that were with triple rank (135 trees/ha) were again thinned, now systematically, leaving 45 trees/ha. Soybeans produced at the same level as the single system in the first three years, while corn only in the first two years. After the first thinning, there was a recovery of grain yield levels for two years for soybean and for one year for corn, in the system with systematic thinning. And after the thinning in 8th year, soybean returned to the levels of the single system as well, at least for 2 years and corn in first year as well. / WCCLF. Evento online.