Abstract Global wetlands are the largest and most uncertain natural source of atmospheric methane (CH4). The FLUXNET-CH4 synthesis initiative has established a global network of flux tower infrastructure, offering valuable data products and fostering a dedicated community for the measurement and analysis of methane flux data. Existing studies using the FLUXNET-CH4 Community Product v1.0 have provided invaluable insights into the drivers of ecosystem-to-regional spatial patterns and daily-to-decadal temporal dynamics in temperate, boreal, and Arctic climate regions. However, as the wetland CH4 monitoring network grows, there is a critical knowledge gap about where new monitoring infrastructure ought to be located to improve understanding of the global wetland CH4 budget. Here we address this gap with a spatial representativeness analysis at existing and hypothetical observation sites, using 16 process-based wetland biogeochemistry models and machine learning. We find that, in addition to eddy covariance monitoring sites, existing chamber sites are important complements, especially over high latitudes and the tropics. Furthermore, expanding the current monitoring network for wetland CH4 emissions should prioritize, first, tropical and second, sub-tropical semi-arid wetland regions. Considering those new hypothetical wetland sites from tropical and semi-arid climate zones could significantly improve global estimates of wetland CH4 emissions and reduce bias by 79% (from 76 to 16 TgCH4 y−1), compared with using solely existing monitoring networks. Our study thus demonstrates an approach for long-term strategic expansion of flux observations.