Cattle ranching is the largest driver of deforestation in the Brazilian Amazon. The rainforest- to-pasture conversion affects the methane cycle in upland soils, changing it from sink to source of atmospheric methane. However, it remains unknown if management practices could reduce the impact of land-use on methane cycling. In this work, we evaluated how pasture management can regulate the soil methane cycle either by maintaining continuous grass coverage on pasture soils, or by liming the soil to amend acidity. Methane fluxes from forest and pasture soils were evaluated in moisture-controlled greenhouse experiments with and without grass cover (Urochloa brizantha cv. Marandu) or liming. In parallel, we assessed changes in the soil microbial community structure of both bare pasture soil as well as rhizosphere soil through high throughput sequencing of the 16S rRNA gene, and quantified the methane cycling microbiota by their respective marker genes related to methane generation (mcrA) or oxidation (pmoA). The experiments used soils from eastern and western Amazonia, and concurrent field studies allowed us to confirm greenhouse data. The presence of a grass cover not only increased methane uptake by up to 35% in pasture soils, but also reduced the abundance of the methane-producing community. In the grass rhizosphere this reduction was up to 10-fold. Methane-producing archaea belonged to the genera Methanosarcina sp., Methanocella sp., Methanobacterium sp., and Rice Cluster I. Further, we showed that liming compromised the capacity of forest and pasture soils to be a sink for methane, and instead converted formerly methane-consuming forest soils to become methane sources in only 40-80 days. Our results demonstrate that pasture management that maintains grass coverage can mitigate soil methane emissions, if compared to a bare pasture soil.