The paper presents the development, implementation and performance investigation via simulations and experiments of a comfort-oriented control strategy for natural ventilation and mechanical air conditioning management of a mixed-mode building. The proposed comfort-oriented control strategy determines whether it would be possible to operate in natural ventilation mode or in mechanical heating/cooling. The control algorithm calculates first the optimal opening percentage of the windows according to adaptive thermal comfort criteria. If natural ventilation cannot guarantee the thermal comfort requirements and mechanical conditioning is required, the algorithm dynamically optimises the heating or cooling set-point targeting a defined Predicted Mean Vote (PMV) index objective. The performance of the proposed controller was tested via simulations and experiments by using a residential mixed-mode building as a case study. The house features operable windows, a reverse-cycle ducted air conditioner and a comprehensive experimental control and monitoring infrastructure. A comparison with a baseline control strategy was performed to evaluate the comfort and energy performance improvement potential of the proposed control algorithm. The comfort-oriented controller was proven to outperform the baseline controller in terms of maintaining comfort in accordance with targets set by the current comfort standards, such as deviation from a PMV set-point or the middle of the adaptive thermal comfort band. The building energy consumption was also reduced in cooling dominated conditions. The experimental tests demonstrated that this logic can be integrated in an embedded controller, and its performance is in line with the expected one from the simulation results.