In semi-arid regions, wetlands are increasingly converted to agriculture use due to growing food demand. Current practices like mowing and grazing considerably alter plant biomass allocation patterns, yet the underlying functional mechanisms remain unclear. To clarify this, we combine community-level leaf and root traits with biomass partitioning analysis across a management gradient (from natural to heavily grazed wetlands) in China's Songnen Plain. Results revealed that agricultural intensification triggered strategic biomass reallocation: aboveground biomass (AGB) and belowground biomass (BGB) decreased 83.1 % and 51.9 % from natural wetlands (NW, 1093.5 g m-2) to heavily grazed sites (HG), while root-to-shoot ratio (RSR) increased 12-fold (HG:3.6 vs. NW:0.3). This coincided with the coordinated leaf-root economic strategies: NW favored acquisitive strategies (e.g. peak high leaf area [LA] =1917.1 mm2, and specific root length [SRL] = 3219.3 cm g-1), whereas mowed/grazed sites optimized conservation investments (e.g. peak leaf carbon concentration [LCC] = 487.4 g kg-1, and root carbon concentration [RCC] = 391.0 g kg-1), illustrating a functional shift from resource acquisition to conservation. Correlation analysis and stepwise regression identified leaf dry weight (LDW), leaf carbon-to-nitrogen ratio, leaf phosphorus concentration, LA, LCC, root length, specific root area (SRA), RCC, and root surface area (RSA) as key predictors for AGB, whereas BGB was predicted by LCC, LDW and leaf nitrogen-tophosphorus ratio, RSA, SRA and root phosphorus concentration (RPC). Leaf traits (LA and LDW) and root traits (RCC, RPC, and SRL) were identified as the best predictors of RSR. Variance partitioning analysis showed that leaf-root trait interactions dominated biomass production and allocation (76.9 %-89.8 % of explained variance), underscoring an integrated root-leaf economic network. Structural equation modeling further revealed that the decreased LA indirectly enhanced RSR through positive effects on AGB and synergistic interactions with LDW, SRL and RCC. Contrasting effects of LA-LDW/SRL interactions on BGB further regulated RSR. LA-driven synergies with SRL and trade-offs between RCC and SRL directly modulated RSR. These results collectively demonstrate that trait interaction networks, rather than individual traits, serve as the key driver of biomass production and allocation in agriculturally managed wetlands. This functional trait-coordination framework advances our understanding of plant strategies for enhancing wetland resilience in semi-arid agricultural landscape.