In this work, we propose a method to couple the behaviour models developed with Python in a previous paper with the dynamic thermal simulation software EnergyPlus, an advanced code used in research and design. The proposed coupling method is applied to the thermal model of an office building situated in the humid tropical climate of Reunion Island after calibrating and validating it with measured temperature and relative humidity data. Then, this resulting coupled model is compared with a typical design office energy model where behaviours are based on typical, deterministic scenarios. The comparison focuses on the power level of the ceiling fans employed, the level of opening use and the calculation time. The results obtained by coupling with the new behavioural models are better than in the conventional deterministic scenarios, providing a more faithful reproduction of user actions in the design phase.