Abstract: Empirical stochastic multi-variable models for prediction of treatment efficiency of wetlands are presented in this article. Wetlands of seven different shapes are visualized using tracer studies. Two different variants of experiments are carried out. Numerous flow rate variations are performed keeping surface area of the wetland constant. The experiment is also carried out with a variation in volume of the wetland which helps to study the effect of flow height on the hydrodynamics within the wetland. A multi-variable model for treatment efficiency in terms of change in tracer concentration as a function of shape, volumetric height of water within the wetland, time, and mass flow rate is considered. Further, another set of experiments is performed studying the treatment efficiency in terms of electro-kinetic parameters. This involves measuring the pH, turbidity, temperature, electrical conductivity, total dissolved salts at inlet and outlet and residence time with varying flow rate, and height of water for the seven different wetland models under study. The electro-kinetic parameters changes due to difference in concentration of the tracer dye which simulates impurities. In this case, treatment efficiency is expressed as a function of the above-discussed electro-kinetic variables, time variation, water height, as well as variation in the mass flow rate. The stochastic multi-parameter models, thus, empirically derived in the above two cases have high coefficient of determination. The models thus derived may be used as a tool for quick analysis of treatment efficiency of any shape and size of a three-dimensional wetland. © 2014 Balaban Desalination Publications. All rights reserved.