Peracetic acid (PAA) is an emerging disinfectant for the treatment of wastewater. While it would be possible to optimize the design of this system using computational fluid dynamics (CFD), the computational intensity would be high. As an alternative, we show that an Artificial Neutral Network (ANN) based metamodel can approximate the CFD solutions over an 11 dimensional performance space (dimensions, hydraulic characteristics, and chemical kinetics). By sampling the design space using a quasi-random sampling technique, a series of CFD simulations of disinfection characteristics of PAA in a wastewater treatment reactor are carried out. After a training process using 40 different CFD runs are completed, the ANN developed can be used to achieve an optimized design of wastewater treatment facilities with minimal total cost and acceptable disinfection performance efficiency.