Abstract: Weed control and biodiversity conservation are both required for sustainable agricultural production. As duckweeds are capable of colonizing open water surfaces in several days on account of their extremely rapid growth rate, introduction of duckweeds to paddy field is considered to inhibit weed growth as a result of shading. To acquire evidence for the utility of duckweed to control weeds and the impact on weed community composition in paddy fields, a field experiment was conducted for a rice-growing seasons. Using community ecology methods, we investigated dynamic changes of weed quantity and community diversity at 4 rice growth stages (tillering stage, booting stage, flowering stage, and maturing stage) and rice grain traits at maturing stage under three treatments: rice grown with Spirodela polyrrhiza mulching (SP), rice grown with Landoltia punctata mulching (LP), and rice grown without duckweed mulching (CK). (Landoltia punctata and Spirodela polyrrhiza were introduced into the respective plots with about 70% floodwater coverage on the day of rice transplantation.) The results showed that: ten species from seven families were recorded in weed surveys. Both duckweed species significantly suppressed weed emergence. SP and LP significantly reduced weed density by 60.3%-75.8% and 81.1%-90.4%, respectively, in the first two stages; and reduced weed biomass by more than 48.0% and 81.3%, respectively, throughout the rice-growing season. Both duckweed species changed the composition of weed communities. Cyperaceae weeds were predominant in the weed communities of the SP and LP treatments. Biomass proportions of Cyperaceae weeds were more than 62.9% and 60.2% in the weed communities of the SP and LP treatments, respectively; while the proportions of broadleaf weeds decreased significantly. Cyperaceae weeds Scirpus juncoides and Cyperus iria showed higher important values in the weed communities of the SP and LP treatments. In the weed communities of the CK, broadleaf weeds Monochoria vaginalis showed the highest important values. The Bray-Curtis similarity indice between weed communities of CK and duckweed mulching treatments was less than 0.25. Correspondingly, the weed communities of CK were separated from the weed communities of duckweed mulching treatments in UPGMA clustering. Duckweed mulching showed no negative impact on plant diversity and rice grain yields. In most rice growth stages, the Margalef index, Shannon-Wiener index, and Simpson index of weed communities under three treatments were not different significantly. Due to the mitigation of weed competition and the potential nutrients supplied by decomposition of duckweed, SP and LP significantly improved grain number per panicle by 33.7% and 15.8%, respectively, and panicle weight by 28.2% and 17.0%, respectively. Rice grain yield in the SP treatment significantly increased by 28.0%. In summary, duckweed mulching reduced weed density and biomass and altered weed community composition while maintained weed diversity in paddy fields, and Spirodela polyrrhiza mulching improved rice yield. The results demonstrated that duckweed introduction to paddy fields is an effective and eco-friendly weed management strategy which is conducive to biodiversity conservation and sustainable agricultural development.