The PID control algorithm applied in the paper is still one of the most popular in many branches of industry. Modern temperature controllers should be equipped with an identification algorithm which allows the determination of the controller settings automatically with high accuracy. This approach leads to the idea of autotuning controllers, which are often based on knowledge of the frequency response of the system. In the paper the multifrequency binary testing identification method in the frequency domain using multifrequency binary sequences (MBS) is applied. This method provides rich information about process dynamics in the form of a few points of the frequency response even though the process is only slightly disturbed. Additionally, the procedures discussed ensure proper choice of the controller structure and its parameters during a short duration identification experiment as well as correction of the errors due to the slow aspects of the furnace dynamics, This allows the MBS method to be applied in the closed loop structure just after the temperature set point has been reached. Start-up identification, which is also described, is a source of a priori knowledge, complementing the MBS method. The presented theory has been verified by experiments. A 20 kW industrial electric resistance furnace was used in the experimental work.