A green roof is an option for improving a building thermal comfort. The investigation is here performed within the specific climate context of Reunion Island in south hemisphere. This type of roof system involves choice difficulties for plant species that are more favorable to establish that comfort. The objective of this work is to simulate the dynamic behavior of this system towards external requests in wet tropical zones and vis-à-vis influences of certain number of physical parameters related to this system. As long as possible, authors used the electrical analogy method to establish a mathematical model associated to the studied system. Based on this model, a Matlab computing code was finalized. Weather data of Reunion Island were used for simulations; the green roof potential and benefit were highlighted by surveying the temperature gain and the heat flux crossing the roof as well as the energy saving performance. Furthermore, the energy consumption being surveyed while doing sensitivity analysis with Fourier Amplitude Sensitivity Test method, the most influential parameters of the model were identified. Full scale experimental results are provided and consisting in monitoring green roof on the top of a public building. According to the results, we can assert that the green roof decreases heat flux entering through the roof during the day and restrains the restoration of accumulated heat at night. Indeed, the support on which the plantation ground bases affects the building thermal insulation. The experimental data are also conducted to prove the effectiveness of thermal insulation by green roofs in reducing temperature in the building between 5°C and 7°C in relation to plants type and the canopy Leaf Area Index (LAI). A comparison of experimental values and model results is done. Among other uses, this code can be used as a tool for choosing the plants and the drain materials to be experimented on the green roof. The results offer hints to optimize the design and thermal performance of extensive green roofs.