In practical engineering design, obstruction baffles are usually used in free-water-surface constructed wetlands (FWS CWs) to cut the width in half and double the length, thus the aspect ratio (length/width) increased fourfold. The length of the obstruction is often set to difference between boundary length and half the boundary width. In order to explore the design method's reasonability and get an optimal value, the tests of FWS CWs under different aspect ratios are needed. In this study, we explored the influence of obstruction length-to-wetland length ratio (OL/WL) and aspect ratio on the hydraulic performance of FWS CWs. The environmental fluid dynamics code (EFDC) was used to establish the hydrodynamic model and water quality model of FWS CWs. Based on the tracer data and pollutant data of 12 FWS CWs, the calibration and verification of model parameters were carried out through sensitivity analysis and manual parameter adjustment. 2 kinds of wetlands named model wetlands and actual wetlands were established. 6 different areas and aspect ratios were set for the model wetlands, and the same as the actual wetlands. Each model wetland was provided with an obstruction, while the actual wetlands had no obstruction. The actual wetland corresponded to the model wetland one by one, that was, the aspect ratio of actual wetland was 4 times of the corresponding model wetland. The purpose of the simulation was to find an appropriate OL/WL that increased the aspect ratio of the model wetland to 4 times by comparing the hydraulic and treatment performance of the model wetlands and actual wetlands. The model wetland was simulated with multiple OL/WL to find the ratio that could make the model wetland and the actual wetland had the closest hydraulic and treatment performance. In addition, the hydraulic index, namely effective volume ratio, short circuit indicator, Morril dispersion index and moment index, were used to evaluate the similarity between the hydraulic performance of the 2 kinds wetlands. Similarly, the removal rates of total nitrogen and total phosphorous were used to evaluate the treatment performance's similarity. The results showed that: 1) The sensitive parameters affecting the hydraulic performance were background horizontal eddy viscosity, dimensionless horizontal momentum diffusion, wall roughness and bottom roughness, and those affecting the treatment performance were maximum nitrification rate, reference temperature for nitrification and constant benthic flux rate of phosphorous. Among the 12 groups, 7 groups were satisfactory or better in hydrodynamic model with correlation coefficient higher than 0.7, Nash-Suttcliffe higher than 0.4, and relative error smaller than 20%; and 10 groups were satisfactory or better in water quality model. The calibrated and verified EFDC model could be used to simulate the hydrodynamic process and purification process of FWS CWs. 2) The 4 hydraulic index had good consistencies. As the increase of OL/WL, effective volume ratio, short circuit indicator and moment index increased and Morril dispersion index decreased. The larger the effective volume ratio, short circuit indicator and moment index would lead to smaller morril dispersion index and better hydraulic performance, which indicated the hydraulic performance increased with the increase of OL/WL. 3) The removal rate of total nitrogen and the phosphorous didn't change greatly as the OL/WL increased. 4) Changing the area had great impact on the treatment performance. 5) When the aspect ratio ranged from 1, 2, 4 to 4, 8, 16, the appropriate OL/WL was 0.675, 0.850, 0.938, which was different from the values of 0.5, 0.75, 0.875 in the traditional test design or engineering application.