Constructed wetlands have come into widespread use as water quality control systems in urban areas. However Published research into their water quality improvement performance has, been sporadic and often contradictory. In order to address this situation, a cooperative study was initiated in 2009 by Wollongong City Council and the University of Wollongong to investigate the pollutant reduction performance of an existing water quality control pond. The pond was monitored for a period between 2009 and 2010 and a unique method for estimating constructed wetland performance was developed to address limitations found in other studies. This method incorporated automated sampling, high temporal resolution monitoring and standard least squares procedures to fit multivariate statistical models to estimate the pollutant reduction performance. The monitoring results were used to calibrate and validate a model which is able to quantitatively assess uncertainty. Results from this study suggest the method applied could be used as a standard method for estimating the pollutant reduction performance of other similar water quality improvement systems.