Magnetic diagnostics for the new generation fusion reactor “ITER” are required to be extremely reliable since they provide measurements essential for reactor operation and protection, plasma control and for measurement of several parameters fundamental to the plasma operation, such as plasma current and shape, disruptions and high frequency macro instabilities. ITER magnetic diagnostics consist of various sets of inductive coils and loops mounted on the inner wall, outside the vacuum vessel and in some of the divertor cassettes [1]. All these probes measure magnetic field or flux variations with respect to time, requiring a precise integration of the signals to recover the absolute values of the field components. They operate in a harsh reactor environment, subjected to nuclear heat loads mainly due to the neutron radiation, generated by the burning plasma. Difficult or impossible access after assembly requires reliability, especially in the area of wiring, connections and vacuum feed-throughs and in choosing margin against radiation damage and extreme transient electrical loads. Additional disturbing effects can arise when both a strong transient magnetic field and thermal gradient occur within the coil structure. All these aspects set a serial of strict design requirements and imply a serious technical challenge. This paper is focused on the design, simulation and optimization of the ITER divertor magnetic tangential coils. The divertor is one of the components exposed to the highest heat load in a fusion reactor, with a surface thermal peak load of 20 MW/m2. About 15 % of the energy produced by fusion reactions is absorbed in the divertor region. The radially-oriented divertor cassettes are exposed to inhomogeneous and time-dependent neutron flux. Six similar divertor cassettes are instrumented for magnetic measurements. Six pairs of equilibrium coils (normal and tangential to the mounting surface) are mounted within each of these cassettes. Of those, pairs near the top region of divertor dome will be exposed to the highest nuclear heating of all magnetic sensors, 2.5 MW/m3. The most critical issue for the divertor coils is to minimise Radiation Induced Thermo-Electric Sensitivity (RITES) [2] and Thermally Induced Electromagnetic Force (TIEMF) [3] by combining a proper choice of conductor with low temperature variation in the coil. Instead of Mineral Insulated Cable (MIC), which was foreseen as the preferred winding for the magnetic coils, a winding made of ceramic-coated steel wire was recently proposed [4]. It is thought that, for this wire, maintaining a temperature variation in the wiring below 10K will be sufficient to allow long-pulse operation. Variations of the divertor coil design have been investigated and simulated with the help of ANSYS programme. The aim was to keep the temperature variation in the winding pack within this limit. The optimisation of the coil, based only on a cooling by conduction was not sufficient to meet the 10 K target. Therefore, an actively water cooled coil was designed and simulated by the CFD code – ANSYS CFX.