Climate simulations (1991-2000) at 60-km and 30-km resolutions using the Regional Environmental Integration Model System (RIEMS2.0) were made to test its ability to simulate the mean temperature and inter-annual variability of temperature in China and to comprehensively evaluate the influence of horizontal resolution on temperature simulation. The results showed that, (1) whether at high or low horizontal resolution, RIEMS2.0 can simulate the spatial distribution of mean temperature to a degree that is similar to observational data results. With an improvement in horizontal resolution, results show a systematic cold bias which is even closer to observational data results. In the winter and annual system, the cold bias is reduced in China and in a majority of the regions, and these reductions are significant (P<0.01). In summer, however, a warm bias around China and in the majority of its regions increased, but the simulation ability was not improved with an increase in resolution. (2) Simulation capabilities for the temperature inter-annual variability were significantly improved with increased resolution. In particular, the inter-annual variability of the annual temperature and winter temperature improved the most, and the ability to simulate the inter-annual variability in the winter temperature and the annual temperature in China, and in most of its regions, were also greatly improved. However, the ability to simulate the inter-annual variability of the summer temperature in China and in most of its regions worsened.