The aim of this study is to implement an unsteady 2D CFD model to predict the time-evolution of greenhouse temperature and humidity distributions all night long in winter, taking account of both radiative transfers and crop interactions with the inside climate. Beyond the classical conservation equations, specific radiative- and crop-submodels were activated. The latter takes account of both the heat and transpiration fluxes induced by the plants and their associated mechanical resistances which depend on the transfers from the plant to the air. Simulations were performed for a clear night on the basis of data collected in winter 2011 inside a 100 m(2) Venlo glasshouse with Impatiens New Guinea crop grown on shelves. The boundary conditions were modified at each time step. Numerical results were validated against data recorded inside the greenhouse: roof temperature, inside air temperature, transpiration flux and air humidity. They highlight the influence of the dynamic boundary conditions on the evolution of the microclimate inside the greenhouse. This study demonstrates the ability of the CFD code to simulate the greenhouse climate evolution with realism. It also suggests a potential exploitation for designing the heating devices in order to optimize the inside climate and energy consumption for various outside conditions.