Abstract Two types of models have been used when it comes to scaling steam-based injection processes from the laboratory to the field: low pressure and high pressure models. The latter, which uses similar fluids and operating conditions as the field reservoirs, is more suited for the steam injection process. Conducting high pressure and high temperature model experiments is very difficult as many variables, such as steam quality, injection rate and pressure need to be controlled all at once and in real time.An advanced automated process control system has been commissioned to overcome this operating complexity in steam injection wells. Steam quality is calculated using a neural network. This network uses an intermediate temperature in the heater and its control signal to assess the steam quality leaving the heater. The production cooling system controls the temperature of the produced fluids at 60oC in order to achieve a significant difference in density between the process fluids, water and heavy oil. As the water and oil densities are known at the controlled temperature, the water cut is determined by measuring the combined density of the produced fluids with a coriolis flow meter.This study has shown that an automated process control system is capable of controlling and optimizing steam injection processes like the steam-assisted gravity drainage process.