Much of the watermelon (Citullus lanatus) cultivation in Brazil is conducted in sandy soil and topographic conditions that favour the erosion process. Therefore, conservation tillage methods are critical for the sustainability of the production chain of this vegetable crop. We studied different tillage methods and cover crops in watermelon cultivation in the area of the reform of degraded pasture. For this purpose, two tillage methods were established as experimental treatments: minimum tillage preparation with subsoiling only, and no tillage. As cover crops white lupine (Lupinus albus) and bristle oat (Avena strigosa) were seeded. As control, watermelon was cultivated with conventional tillage, without prior cultivation of cover crops. For the experimental design, randomised blocks in a factorial arrangement with four replications were used. After liming and phosphate fertilisation of the soil, cover crops were cultivated in soil with minimum tillage and no tillage to produce straw to be used for soil cover, where subsequently the watermelon was grown. The productivity of dry mass and nutrient accumulation in the shoot of cover crops, the soil properties and the watermelon agronomic performance were evaluated. White lupine had better performance in the production of dry mass and nutrient accumulation in shoot than bristle oat. There were differences among treatments for soil penetration resistance, where in conventional tillage the values were lower in the first 30 cm of depth in relation to no-tillage cultivation. The tillage method also affected the fertility of the soil at a depth of 0 to 20 cm. The no tillage provided increased nitrogen leaf content in watermelon regardless of cover crops but restricted root growth in relation to minimum tillage and conventional tillage. Watermelon had similar commercial production by different treatments, with reduction only in no tillage on bristle oat straw.