The objective of present study was to assess the water quality parameters in two experiments, comprising six configurations of pilot-scale vertical subsurface flow constructed wetlands (VSFCWs) by the comparison of the removal efficiency of organic matter and nutrient pollutants from water in arid region. Effluent treatments were studied in 5-month experiment under different operational conditions including (1) substrate type: sand (S) (experiment 1) or fine gravel (G) (experiment 2) and (2) agronomic species: Phragmites australis (Ph) or Typha latifolia (Ty). This experiment demonstrated that the vegetated wetlands were more efficient than non-planted (Np) ones in terms of removal of TSS, COD, NO3-N, and TKN from wastewater in the first experiment and more efficient for all water quality parameters except for TSS and BOD5 in the second experiment. For a given species, the efficiency of both experiments was generally higher with Ph-S than Ph-G and with Ty-S than Ty-G. Regarding unplanted CWs, Np-S performed better than Np-G for all of the tested water quality parameters except TSS. The wetland efficiency indicated that CWs planted with P. australis contributed greatly to the removal of COD, NO3-N, and TKN in the first experiment, and NO3-N, NH4-N, PO43-P, and TP in the second experiment, whereas CWs planted with T. latifolia provided the highest removal only with TSS for the first experiment, and COD and TKN for the second experiment. The highest efficiency of unplanted setups (Np-S) was for BOD5, NH4-N, PO43-P, and TP, while with Np-G was only for TSS and BOD5. In general, the first experiment is better than the second in removal efficiency for most of the tested parameters. Therefore, the use of sand substrate was more suitable than gravel for wastewater treatment in VSSFCW. As well as, P. australis performed better than T. latifolia for most of the studied parameters.