Abstract: Agricultural runoff generally has a low carbon-to-nitrogen (C/N) ratio, and efficient nitrogen removal in constructed wetlands (CWs) remains challenging when treating this type of wastewater. Using biochar, FeCl₃-modified biochar, and iron ore as substrates has great potential to enhance total nitrogen (TN) removal efficiency in CWs. However, the underlying microbial mechanisms remain unclear. Five CWs with different substrates, quartz sand constructed wetland, biochar constructed wetland, FeCl₃-modified biochar constructed wetland, iron ore constructed wetland, and iron ore + biochar constructed wetland, were built to compare their treatment performance. The rhizosphere microbial community compositions were analyzed to reveal the underlying mechanisms driving their treatment performance. The results showed that iron ore was the most efficient amendment, although all treatments increased TN removal efficiency in the CWs. The high porosity and large specific surface area of biochar made it an excellent adsorbent for removing \mathrmNH_4^+ -N from the influent, which, in turn, increased TN removal efficiency in CWs. FeCl₃-modified biochar application increased TN removal efficiency in CWs mainly by stimulating the relative abundance of denitrifying bacteria such as Actinobacteria, Pseudomonas, and Flavobacterium. Iron ore amendment increased TN removal efficiency primarily by promoting the relative abundance of Geobacter (Fe-ammox bacteria) and nitrate-dependent anaerobic ferrous oxidizing (NAFO) bacteria such as Proteobacteria, Bacteroidetes, and Dechloromonas. Iron ore + biochar addition increased TN removal efficiency largely through improving the relative abundance of denitrifying bacteria such as Firmicutes and Sphingobium. The synergistic effect of Fe-ammox bacteria and NAFO bacteria in iron ore CWs resulted in the highest TN removal efficiency for agricultural runoff with low C/N. The above results suggested that it was a feasible way to apply iron ore to treat agricultural runoff with low C/N.