Abstract: In recent years, the meat duck industry in China is mainly dominated by small-scale free rearing. This single-layered flat rearing pattern was characterized by low density, difficulty in feces and sewage management and simple infrastructure, which was weeded out due to the intense pressure for land resources and environment protection. Meanwhile, there was an increasing scale in the development of meat duck industry and an increasing tendency in the spatial aggregation. The cascading cage-rearing laying of meat duck is the rapidly developed stereoscopic breeding pattern. Compared to the traditional duck-rearing pattern, the new meat duck rearing system has an advantage in the improved land use rate, convenient duck feces processing and reduced labor costs. However, there is an accumulation of hazardous gas, particulate matter and airborne microorganism in intensive poultry houses. Poor ventilation not only affects the air quality in poultry houses, but also significantly increases the mirco-environmental humidity, thereby degrading animal's health and growth. Therefore, the reasonable design and well airflow organization is an important measures for the improvement of environmental quality in houses. Duck house were built at Xintai Ecological Meat Duck Research Base of Shangdong Province (35°30′N, 117°27′E). Length, width and ridge height of the meat duck house were 80.0 m, 13.0 m and 4.0 m. Roof height was 2.6 m. Upper architectural system of the duck house was light steel, while the wall body was a brick-cement structure. 24 ventilation window (length×height=560 mm× 360 mm) were each installed in the south and north walls of duck cages. These ventilation window were evenly distributed at the upper parts of the south and north walls. The ventilation window was controlled by 100 mm-thick color steel polystyrene sandwich boards. Lower parts of the ventilation window were 1.94 m to the floor. Three layers × five columns of cascading individually cages (Length × Width × Height= 69.8 m×1.0 m×2.02 m) were installed inside the duck house. There were 67 duck cages at each layer. Fifteen Cherry Valley meat ducks were reared in an individual cage which size was 1 000 mm × 1 000 mm × 400 mm (Length × Width × Height). To construct the CFD model of duck house, temperature, humidity, CO2 and NH3 concentration in the cascading duck-rearing cages were measured. Temperature field and airflow field of the duck house were simulated and an optimal design for the air conditioning pattern of the present duck house was raised. Results of the study showed that during the test period air temperature in the duck house was 11.35-20.68 ℃, while the average temperature difference between the inner and outer side of the duck house was 10.86 ℃. The average relative humidity inside the duck house was 70.27%, slightly higher than that outside the house. Average concentrations of CO2 and NH3 in the house were 3 285 and 0.33 mg/m3, respectively. The average relative error between the simulated temperature field and airflow field of the original duck house and the measured results was 7.21%-8.87%, suggesting a very good consistence between the digital simulation of the constructed model and the experimental results. Air conditioning was further optimized by lowering the height of ventilation window in either wall. Results of simulation showed that the temperature range of inside the house after model optimization was 11.07-19.71℃. The fluctuation range of temperature was dramatically reduced compared to the original model, which had reduced the temperature-induced stress for ducks. The average air speeds of optimized model W1, W2 and W3 were 0.34, 0.34 and 0.31 m/s, respectively. This was more even than the original model. Results of the present study will supply theoretical evidence to the reasonable design and improvement of the environmental quality for the similar kind of duck house.