The paper presents techniques for the design and training of Artificial Neural Networks (ANN) models for the dynamic simulation of the controlled Benchmark Simulation Model no. 1 (BSM1) Waste Water Treatment Plant (WWTP). The developed ANN model of the WWTP and its associated control system is used for the assessment of the plant behaviour in integrated urban waste water system simulations. Both embedded PID (Proportional-Integral-Derivative) control and Model Predictive Control (MPC) structures for the WWTP are investigated. The control of the Dissolved Oxygen (DO) mass concentration in the aerated reactors and nitrate (NO) mass concentration in the anoxic compartments are presented. The ANN based simulators reveal good accuracy for predicting important process variables and an important reduction of the simulation time, compared to the first principle WWTP simulator.