The practical implementation of a control process to control the blood sugar level of a dog is discussed. The problem was to maintain the blood sugar level of a dog at the dog's fasting level inspite of a disturbance due to a constant infusion of insulin. The process control consists of two parts; an identification system used to estimate the process parameters, and an adaptive control strategy to control the process based on the estimated parameters. The identification system used was ordinary least squares, and the control strategy was proportional-integral derivative control. The effects of filtering on the above control strategy was also examined. Simulation results as well as results of on-line experiments are presented.