Water quality and its effects on human life have become one of the major concerns in aquatic ecosystems. The water quality index (WQI) is defined as a parameter to interpret water-monitoring data and clarify the quality of water. In this study, the gene expression programming (GEP) and artificial neural networks (ANNs) were employed to predict WQI in free surface constructed wetlands. Seventeen points of a selected wetland were monitored twice a month over a period of 14 months, and an extensive data set was collected for 11 water quality variables (WQVs). A principal factor analysis (PFA) indicated that WQI was greatly affected by pH and SS, while temperature no has significant effect on the WQI in tropical areas. A sensitivity analysis was carried out to reduce the number of 11 WQVs in prediction of the WQI. Subsequently, five significant parameters, pH, suspended solid (SS), ammoniacal nitrogen (AN), dissolved oxygen (DO) and chemical oxygen demand were selected to develop a GEP and ANNs. The GEP was able to successfully predict the WQI with high accuracy (R 2 = 0.983 and MAE = 0.295). The statistical parameters indicate that, although the ANNs with R 2 = 0.988 and MAE = 0.013 produced better results compared with GEP, the GEP-based formula is more useful for practical purposes. The GEP and ANNs are recommended as rapid and powerful WQI evaluation techniques to reduce substantial effort and time by optimizing the calculations.