Abstract: As one of the most effective cooling method, the fan-pad evaporative cooling system has been widely used to provide a suitable growth environment for greenhouse crops. An optimization method of the fan-pad cooling system based on computational fluid dynamics (CFD) was proposed to improve the cooling performance inside the greenhouse in summer. The Reynolds-averaged Navier-Stokes equations were solved using finite volume method (FVM). Due to the remarkable effect of gravitation on the microclimate distribution inside the greenhouse, the Boussinesq hypothesis was taken into account. The standard k-ε turbulent model was selected to predict the distribution of air flow. Solar ray tracing was applied to load the solar radiation model, while the discrete ordinate model was selected for considering the effect of thermal radiation. Crops in the greenhouse were regarded as the porous medium, which was governed by the Darcy-Forcheimier equation in the CFD model. A three-dimension greenhouse model was developed to simulate the microclimate distribution and air circulation inside the greenhouse adopting fan-pad cooling system. The verification experiment was conducted in a Venlo-type greenhouse in the campus of Zhejiang University of Technology (30°14′N, 120°09′E) from 12:30 to 13:30 on July 23, 2012. Thirteen observation points of T1-T10 and TH1-TH3 were set up in the experimental greenhouse to validate the simulated air temperature and velocity. The errors between simulated and measured air temperature at the observation points varied from 0.7 to 2℃, and the errors of air velocity were less than 0.13 m/s. Compared with the measured values, the absolute mean errors of simulated temperature and air velocity were less than 4% and 6% respectively. It proved that the CFD method is reliable to estimate the distribution of air velocity and temperature in the greenhouse. The validated CFD model was then used to further analyze the cooling performance of different greenhouse cases in terms of the greenhouse lengths, the evaporative pad areas and the greenhouse ventilation rates. The indoor environment with the temperature of below 30℃ and the velocity of below 1 m/s was suitable for crop growth, and this condition was used as a criterion for optimal design. Based on the orthogonal test method, greenhouse cases with different greenhouse lengths, evaporative pad areas and air velocities of fans were classified and simulated to analyze their relations. The simulations illustrated that the greenhouse ventilation rate of 153.1 m3/(m2·h) and the minimum pad area of 6 m2 can meet the cooling requirement in a Venlo-type greenhouse with 24 m length and 9.6 m width. In contrast with greenhouse of 70 m length, the maximum pad area of 13.5 m2 had to be configured, because the greenhouse with smaller evaporative pad need combine with the fan's velocity of more than 105 m3/(m2·h). According to the relations among greenhouse length, evaporative pad area and fan's velocity resulted from CFD analysis; the fitted results could be achieved to design the fan-pad evaporative cooling system in the greenhouse in eastern China. The fitting optimization showed good agreement with the previous corresponding research results, which demonstrated that CFD technique was rational and reliable to design the fan-pad evaporative cooling system in the greenhouse.