One of the main critical points in the production process of orange trees is the negative interference imposed by weeds. Considering the relevance of this factor, this study aims to evaluate integrated methods for weed control on the yield of orange (Citrus sinensis L.), cv. 'Natal'. A field experiment was carried out in a randomized block design with the following nine treatments: manual weeding of the area associated with maintenance with herbicide; herbicide application in the total area; herbicide application in the total area followed by sowing of jack bean (Canavalia ensiformis L.) between the rows; herbicide application in the total area followed by sowing of the legume in the total area; herbicide application between the rows only in the free space between canopies; herbicide application in the row; herbicide application in the canopy projection; herbicide application in the total area except for the canopy projection; besides a control treatment, without weeding. The herbicide used was glyphosate, always applied at a dose of 1,080 g ha -1 . Among the studied managements, it can be concluded that the manual weeding of the area associated with maintenance with glyphosate (1,080 g ha -1 ); glyphosate application in the total area; glyphosate application in the total area followed by sowing of jack bean between the rows; glyphosate application in the total area followed by sowing of jack beans in the total area; and glyphosate application in the canopy projection resulted in the same yield as the traditional management adopted by the farmer, which consists of glyphosate application in the row. On the other hand, glyphosate application in the total area except for the canopy projection and glyphosate application between the rows only in the free space between canopies were not good weed control methods, resulting in lower yield, whose value was similar to that of the control without weeding (in the bush).