Due to the efficient combination of a forming step with a consecutive heat treatment, hot-stamping has become an established technology for the production of high strength steel parts in the automotive industry. In the beginning, sheets are heated above austenitisation temperature and held in order to obtain a fully austenised microstructure, then formed and instantly quenched in the forming tool. To achieve the desired increase in tensile strength for the widely used manganese boron steel 22MnB5, cooling rates of at least 27 K/s are necessary. This requirement sets a high demand on the numerical process simulation in order to being able to predict the occurrence of component or process errors with a high degree of certainty. To achieve this, the exact knowledge of the local heat transfer coefficient is necessary, which dominantly determines the temperature distribution within the work piece and the die. Since there is none standardised test method for the determination of heat transfer coefficients exists, a practical test method is presented in this study. In addition to the use of a divisible temperature-measuring stamp, the method is based on a close coupling of practical experiment and iterative numerical simulation. With the method and tools shown in the scope of this paper, the heat transfer coefficient could be successfully determined as a function of contact pressure and tool start temperature, taking the process route of hot-stamping into account. Results are compared with literature knowledgeorder to demonstrate the performance of the determination method.