Aquatic protected areas and co‐management have been proposed as strategies for conserving fisheries resources in several world regions. We analyzed the effects of different levels of governance on yield of experimental fisheries in Amazonian floodplain lakes. Sampling between 2018 and 2022 used gill nets deployed for 24 h during flooding and receding water phases. Generalized linear models (GLMs) were fitted to catch per unit effort (CPUE) and covariates, river level, time, sampling sites, level of management of the collection sites, and lake fishery management type. CPUE was significantly related to governance, management type, sampling site, and water level. Our results indicated that areas of sustainable use where community‐based co‐management was implemented were more effective for promoting high fishery yield.