Traditional constructed wetland designs typically result in variable efficiencies for trace organic contaminant removal. In this work, we used a Box–Behnken experimental design for optimizing the conditions of pH, nitrate concentration, and dissolved organic carbon (DOC) concentration that would maximize the rate of triclosan phototransformation while minimizing the accumulation of toxic byproducts. Triclosan is a frequently detected and toxic antimicrobial agent present in many consumer and industrial products. The results showed that high pH values (9.9) and low DOC concentration (11 mg/L–) would maximize triclosan phototransformation rate while minimizing the accumulation of toxic byproducts. As long as DOC concentrations were larger than 33 mg/L, nitrate concentration did not show a significant effect on triclosan phototransformation rate. The major transformation products detected were 2,4-dichlorophenol and compounds with chemical formulas C12H8Cl2O2 and C12H9ClO3, resulting from a chlorine loss or replacement by a OH group. In addition, 4-chlorocatechol was mainly detected during direct photolysis and 2,8-dichlorodibenzo-p-dioxin was only found during direct photolysis at pH 8. This study showed that wetland efficiency at removing triclosan can theoretically be increased by limiting DOC-contributing factors, e.g., emergent vegetation, and supporting pH-increase processes, e.g., via algae growth or by incorporating alkaline geomedia.