The traditional approach for vehicle thermal development relies heavily on experimentation and experience. A virtual vehicle would be very beneficial in providing upfront engineering support which should lead to time and cost savings. To realize a useful model, the authors have based their approach on experimental data and correlations for each significant vehicle component. The vehicle has been divided into five linked modules representing powertrain cooling and cab climate. The paper describes the approach taken for each module and shows that good agreement can be found between model predictions and actual measurements.