Abstract Background: The use of pre-emergent herbicides is an important tool to control weeds, however the tolerance of soybean to these herbicides can vary according to the type of soil. Objective: The aim of this study was to evaluate sulfentrazone and diclosulam soybean selectivity in a sandy soil, in order to establish these herbicides as supporting tools in weed control. Methods: The treatments consisted of six doses of sulfentrazone (150 to 400 g ha-1) and diclosulam (25.2 to 75.6 g ha-1), and an untreated control. A field study was repeated in two growing seasons (2013/14 and 2014/15) in a sandy soil. Results: The occurrence of injury depended on the growing season. The first season presented lower rainfall rates during the crop cycle. The recommended dose of sulfentrazone (200 g ha-1) caused 26% and 10% of plant injury at 15 days after the treatment for the first and second season, respectively. For diclosulam, the recommended dose of 35.3 g ha-1 caused 20% and 8% of plant injury, respectively, for the first and second season. However, at the recommended doses, only for diclosulam and in the rainy season there was a reduction in soybean productivity. Conclusions: The herbicide sulfentrazone, although causing visual soybean plant injury, present satisfactory selectivity at recommended doses and can be used as an important tool on weed control on sandy soils. Similar response was observed for diclosulam in 2013/14. However, in a rainfall season diclosulam impacted on soybean yield at the recommended dose for sandy soils, with selectivity depending of the growth season.