ABSTRACT The objective was to study the selectivity of herbicides bentazon and nicosulfuron, alone and in a mixture, to Crotalaria juncea, when established in intercropping with maize, as well as weed control and maize development in this production system. Two experiments were conducted in the field, in late harvest and in harvest periods. The experimental design was a randomized block, with 14 treatments and four replicates. Eight treatments with herbicides bentazon (720 and 960 g a.i. ha-1), nicosulfuron (16 and 48 g a.i. ha-1) and bentazon plus nicosulfuron (720 + 16; 720 + 48; 960 + 16; 960 + 48 g a.i. ha-1) were studied, besides six controls (single crop maize and sunn hemp, with and without weed control, intercropped maize and sunn hemp, with and without weed control). Herbicide bentazon alone, in both dosages evaluated, was selective to C. juncea. Nicosulfuron in the dosage of 48 g a.i. ha-1, alone or in a mixture with bentazon, promoted the best weed control. Maize interferes in the recoverability of sunn hemp plants treated with herbicides. C. juncea intercropped without herbicides or sprayed with bentazon negatively impacted maize yield in the harvest. The pulverization of bentazon (720 and 960 g a.i. ha-1) in late harvest and bentazon plus nicosulfuron (720 +16 and 920 + 16 g a.i. ha-1) in harvest provided the best results, considering maize yield, development and recovery of sunn hemp and weed management.