The ability of German regional climate Model (REMO) to simulate the near surface air temperature and total precipitation over China in 1989-2008 were assessed with the use of Taylor diagrams and bias analysis. Comparing the simulated near surface air temperature with a 20-year observational dataset from China, the spatial correlation coefficient was relatively high (0.94). However, the spatial correlation coefficient for total precipitation was relatively low (about 0.42). The near surface air temperature simulated by REMO was higher than the observed values in most part of China, showing a bias range within 4 C. Significant cold bias of about -4 to -2 C occurred over most of the Qinghai-Tibetan Plateau. In terms of total precipitation, the simulated values were higher than the observed ones, with biases evenly distributed. The annual mean bias in most part of China was within 300 mm. Except for the Qinghai-Tibetan Plateau, South China and Southwest China, REMO accurately reflected the distribution of near surface air temperature and total precipitation. REMO represented the temperature and total precipitation well in North China and Northeast China. REMO simulations were quite close to observations for near surface air temperature in summer and total precipitation in winter. REMO still needs to be improved in complex terrain areas.