Rice sustains the survival of nearly half of humanity, meanwhile, paddy systems are also a significant source of greenhouse gas (GHG) emissions in agriculture. Drainage is widely accepted as a key agricultural management practice (AMP) to reduce CH 4 emissions in paddy systems, however, its impact on rice yield and global warming potential (GWP) under different soil and APM conditions remains uncertain. Here, we performed a global meta-analysis of 519 samples from 74 fields in 64 peer-reviewed publications, including measurements of rice yield and at least one of GHG emissions (N 2 O, CH 4, and CO 2). Overall, the effects of implementing drainage in paddy systems on yield, N 2 O emissions, CH 4 emissions, CO 2 emissions, GWP, and GWP intensity (GWPI) were+ 0.3%(p= 0.85),+ 149.9%(p< 0.0001),− 57.8%(p< 0.0001),+ 27.7%(p= 0.12),− 44.9%(p< 0.0001), and 57.7%(p< 0.0001), respectively. Drainage in lower total N soils was more capable of maintaining yield (+ 0.9% vs.− 1.4%), and drainage in higher pH and organic carbon soils resulted in greater reductions in CH 4 emissions (− 46.73% vs.− 37.3%,− 46.36% vs.− 38.12%). Mid-season drainage was superior to early-season and fallow season drainage in CH 4 emissions reduction. Extending the duration and increasing the events of drainage would increase N 2 O emissions but contribute to a better CH 4 emissions reduction effect. Implementing drainage in paddy systems with organic amendment can achieve greater GHG emissions reduction (both N 2 O and CH 4) than in that without organic amendment. The GWPI decreased by the installation of drainage system might disappear when the nitrogen fertilizer application rate (N rate) reached over 530 kg ha− 1. As for different N rates, single drainage is more suitable for the N rate below 115 kg ha− 1, while multi-drainage has more advantages on reducing GWPI in the high N rate (≥ 115 kg ha− 1).