In milling machines, waste heat from motors, friction effects on guides and most importantly the milling process itself greatly affect positioning accuracy and thus production quality. Therefore, active cooling and lead time are used to reach thermal stability. A cheaper and more energy-efficient approach is to gather sensor data from the machine tool to predict and correct the resulting tool center point displacement. Two such approaches are the characteristic diagram based and the structure model based correction algorithms which are briefly introduced in this paper. Both principles have never been directly compared on the same demonstration machine, under the equal environmental conditions and with the same measurement setup. The paper accomplishes this comparison on a hexapod kinematics examined in a thermal chamber, where the effectiveness of both approaches is measured and the strengths and weaknesses of both are pointed out.