Abstract: The objective of this study was to evaluate the validity and suitability of denitrification-decomposition model(DNDC)to estimate grain yield and ammonia volatilization in double rice cropping system. The different nitrogen application levels experiments were conducted under double rice cropping system, located in the middle and lower reaches of the Yangtze Plain. A continuous airflow enclosure (CAE) method was used to estimate NH3 emission, and DNDC model was used to simulate grain yield and ammonia volatilization. The uncertainty and sensitivity analysis was conducted to assess the variability in the model outputs and input parameters of climate conditions, soil properties and farming management etc. The results showed that the DNDC simulated grain yield and ammonia volatilization fitted well with the field observed data. The model can capture the peak of ammonia volatilization caused by fertilizer N application. The correlation coefficients between simulated and observed grain yields of early rice, late rice and double rice cropping system were 0.994, 0.928 and 0.979, respectively. The correlation coefficients between simulated and observed values of ammonia volatilization in early rice, late rice and double rice growth period were 0.994, 0.998 and 0.997, respectively. According to the sensitivity test on the DNDC model, grain yield was significantly affected by temperature, and ammonia volatilization was increased with the nitrogen application rate and temperature. The relationship between grain yield and nitrogen application amount was curvilinear. The nitrogen fertilizer requirement of theoretical maximum yield calculated from DNDC simulation data and field observation data were 420 and 417 kg·hm^-2, respectively. Soil ammonia volatilization increased with nitrogen application rate following a quadratic function equation, rather than a linear equation. A higher reliability was found between DNDC simulated value and field measured results about grain yields than ammonia volatilization. However, there is a certain deviation between the simulated and measured values of ammonia flux and cumulative ammonia volatilization, ongoing modification and calibration still need in order to improve the model's performance. Environmental data such as meteorology, soil, water, etc., as more detailed, accurate, and informative as possible, is necessary to reduce the uncertainty of input data, before DNDC application.