Abstract: Abstract: Goose is seasonal breeding animal, the goslings and adult geese supply show seasonal changes, so that the sale price of the geese is fluctuating. The out-of-season breeding technology was employed to balance annual goose production and achieve higher economic benefits by regulating the normal breeding of geese in the spring and summer of non-breeding season. This technique has been popularized in the last few years, however, the high temperature in the summer is adverse to the goose production in the process of out-of-season breeding. It is known that the quality of thermal environment and the concentration of harmful gases are important factors to the health of livestock. In order to reduce or eliminate adverse effects of heat stress and lack of water on the geese during the process of out-of-season breeding, the ventilation-cooling mode with fans and evaporative pad was employed in conventional goose house, while a large amount of airflow diffusion located in the upper part of the goose house due to that these airflow of small density flowed upward. The use of computational fluid dynamics (CFD) techniques to solve complex fluid problems has greatly increased in the last few years. In this study, the goose house with mechanical ventilation had a large number of airflow diffusion in the upper part of the goose house, this may cause a problem that the ventilation effect of the ground geese was obstructed. So a structure optimization scheme based on computational fluid dynamics (CFD) has been proposed, multiple stretching film structures were installed below the girders of goose house to change the airflow directions as well as to increase the airflow flux in the lower part of goose house around the geese. Based on that, the two factors of vertical angle between stretching film and girder, drop height of stretching film were investigated to obtain uniform airflow and higher ventilation efficiency in the goose house. A three-dimensional steady goose house model was developed by the field measured boundary conditions data. Comparison between simulations and measurements for the 40 test points of wind speed showed that the RMSE, maximum absolute error and average relative error was 0.152 m/s, 0.29 m/s and 2.04%, respectively. It proved that the CFD method is reliable to estimate the distribution of air velocity in the goose house. The validated CFD model was then used to get the optimal combination scheme of 27 different construction cases: In the 42-meter-long goose house, we find that when 10 roll films are installed in the shed, and the angle between roll film and the vertical direction of girder is 60? with maximum drop height of 1.2 meters, the ventilation in the shed has the highest efficiency and its air distribution is more homogeneous, and this simulation is concluded to be the optimal one.Through in-situ test, 40 points are compared before and after the transformation of the wind speed value, the average airflow velocity at the height of 0.5 meters above the ground is up to 1.01 m/s, and the average airflow velocity in the conventional goose house without film is only 0.483 m/s. The test results show that: after transforming averaged wind velocity increased by 0.527 m/s, and the airflow uniformity coefficient decreased by 32.2%, the structure of the film increases the airflow velocity in the lower position significantly. The results of this experiment provide a reference for structure optimization of similar poultry house, the internal environment regulation and the structure design of goose house.