The presented paper focuses on the computation of heat transfer related to continuously variable transmissions (CVTs). High temperatures are critical for the highly loaded rubber belts and reduce their lifetime significantly. Hence, a sufficient cooling system is inevitable. A numerical tool which is capable of predicting surface heat transfer and maximum temperatures is of high importance for concept design studies. Computational Fluid Dynamics (CFD) is a suitable method to carry out this task.In this work, a time efficient and accurate simulation strategy is developed to model the complexity of a CVT. The validity of the technique used is underlined by field measurements. Tests have been carried out on a snowmobile CVT, where component temperatures, air temperatures in the CVT vicinity and engine data have been monitored. A corresponding CAD model has been created and the boundary conditions were set according to the testing conditions. In a first step a simplified study is presented, to gain basic knowledge about the system, followed by a full underhood airflow simulation. The modelling process is presented in detail and necessary adaptions are identified.The results show, that the numerical model is able to predict the surface temperatures within a range of 5 % for different load cases. Thus, the developed method is validated and can be used for future development processes. The influence of the pulley design on the underhood airflow will be evaluated to identify optimization potential. Moreover, the presented work is the first of its kind, where a numerical heat transfer simulation of a CVT is compared to field tests carried out on snow.