The methodology integrating crop models and global climate model are important tools to estimate the impacts of climate change on agricultural production. The characteristics of changes on agricultural climate resources and the impact of climate change on rice production in Zhejiang Province were explored with ensemble simulation of 11 global climate models (GCMs) under 3 representative concentration pathway (RCP) scenarios from the Coupled Model Inter-comparison Project Phase 5 and the ground observation data. Climate projections for 2050s period (2041-2060) and 2070s period (2061-2080) under RCP2.6, RCP4.5, RCP8.5 scenarios were stochastically downscaled to local-scale ensemble daily weather data using the LARS-WG weather generator. The local ensemble weather data were applied as input in rice simulation models ORYZA2000. It was shown that the mean temperature for Hangzhou and Jinhua during 2070s would increase by 1.65-3.56, 1.75-3.67 ℃, respectively, and the rice heat damage would become more serious. Ensemble simulation result indicated that the growth duration of rice in representative site would be shortened due to the accelerated rice development under climate warming. Without direct effect of CO_2 elevation, the rice yield would decrease as compared to the baseline, and higher yield decrease was simulated under high RCP climate scenarios with greater temperature increase.