Antibiotic resistance genes (ARGs) commonly detected in wastewater can potentially lead to a health crisis. Constructed wetlands (CWs) remove ARGs and sulfonamides (SAs) from wastewater, but the importance of plants in the process is seldom reported. We compared the effect of three wetland plant species (Cyperus alternifolius, Juncus effuses, and Cyperus papyrus), sample distance from the root, and SA presence on the environmental abundance of class 1 integron (intI1) and SA resistance genes (sul) using specially designed CW rhizoboxes. Quantitative polymerase chain reaction revealed that the relative abundance of the target genes in planted CWs, especially in C. alternifolius planted CWs, was significantly lower than that in unplanted CWs (P < 0.05). The substrate in the rhizosphere or near-/moderate-rhizosphere (closest to the root) showed the lowest average relative abundance of the target genes, while the bulk substrate (without the root) showed the highest abundance of these genes, irrespective of the planted species. Further, the influence of plants was more evident after 8 weeks of wastewater treatment. The trend was the same in SA-treated and untreated groups, although the relative abundance of the target genes was significantly higher in the former (P < 0.05). The weaker correlation between the intI1 and sul genes in the rhizosphere and near-/moderate-rhizosphere in comparison to the bulk substrate in the SA group suggested that the risk of horizontal gene transfer was probably higher in the bulk substrate and unplanted CW. A partial least-squares path model revealed that dissolved organic carbon and oxygen content significantly influenced SA concentration, microbial community, and intI1 genes, and then shaping the sul genes together. Finally, redundancy analysis suggested that abundance of sul genes was influenced by bacteria enriched in the bulk substrate and unplanted CWs. The findings provide new insights into the importance for controlling risk of ARGs by wetland plants.