Based on observational data of CN05 (daily observations on a 0.5° latitude-longitude grid over China) and outputs of 45 CMIP5 (Coupled Model Inter-comparison Project Phase 5) models adopted in the Fifth Assessment of the Intergovernmental Panel on Climate Change (IPCC AR5), the capability of new generation climate models on simulating air temperature and precipitation over the three provinces in Northeast China during 1961-2005 are evaluated. Results show that: 1) Most of the models are capable of reproducing the significant warming trend during 1961-2005 in the three provinces in Northeast China; however, they have limited ability to realistically simulate the interannual variation of mean surface air temperature. 2) All models can well capture the spatial distribution of surface air temperature, with the MME (the multi-model ensemble mean) more consistent with observations than results of most individual models (the spatial correlation coefficient between MME and observations is up to 0.96). 3) There are large differences in precipitation simulations between various climate models. Although the multi-model ensemble approach has limited ability in simulating the interannual variation and linear trend of regional mean precipitation, it can better capture the spatial pattern of precipitation than any individual models (the spatial correlation coefficient is up to 0.86). Generally speaking, the multi-model ensemble approach has proven effective at simulating the spatial-temporal variations of surface climate in Northeast China. Specifically, it performs better in the simulation of surface air temperature than the simulation of precipitation. The spatial patterns of surface air temperature and precipitation are also better represented than their temporal variations in the model results.