Abstract Today’s ECUs (Electronic Control Units) face a high-power density mounted in a very small volume due to miniaturisation, available space, and weight limitations. For this reason, the resulting heat dissipation has become increasingly difficult to manage and continuous efforts are being made to optimize thermal management in this direction. This paper shows an improvement in heat dissipation by finding an optimal configuration between the fan speed and the distance between the cooling fins using the Design for Six Sigma methods. The focus was on finding a reliable transfer function that could predict the influence of the identified critical design parameters. Two transfer functions were derived and analysed in parallel; one based on Design of Experiments (DoE) and another based on simplified thermal theory. The accuracy of the transfer function and predictions were compared with the results of thermal simulations performed in the Ansys Icepack program. This shows that the use of the DoE method based on a small number of simulations is sufficient to estimate the behaviour of the ECU, without using a more complex theoretical approach.