AbstractSuperconducting magnets for fusion are subjected to large variations of heat loads due to cycling operation of tokamaks. The cryogenic system shall operate smoothly to extract the pulsed heat loads by circulating supercritical helium into the coils and structures. However the value of the total heat loads and its temporal variation are not known before the plasma scenario starts. A real-time heat load estimator is of interest for the process control of the cryogenic system in order to anticipate the arrival of pulsed heat loads to the refrigerator and finally to optimize the operation of the cryogenic system. The large variation of the thermal loads affects the physical parameters of the supercritical helium loop (pressure, temperature, mass flow) so those signals can be used for calculating instantaneously the loads deposited into the loop. The methodology and algorithm are addressed in the article for estimating the heat load deposition before it reaches the refrigerator. The CEA patented process control has been implemented in a Programmable Logic Controller (PLC) and has been successfully validated on the HELIOS test facility at CEA Grenoble.This heat load estimator is complementary to pulsed load smoothing strategies providing an estimation of the optimized refrigeration power. It can also effectively improve the process control during the transient between different operating modes by adjusting the refrigeration power to the need. This way, the heat load estimator participates to the safe operation of the cryogenic system.