Abstract: In this paper，the modeling results of the past 1000 years' climate changes with models MPM­2 and ECHO­G were reported，these climate changes and their relations to forcing factors are analyzed，and the cause and driving factors to result in the climate changes were discussed. Model MPM­2 is an Earth­system model with intermediate complexity，whereas model ECHO­G is a coupled global climate model. The scenarios simulated mainly involve the following forcing factors: solar radiation，volcanic dust，and CO2 concentration. The effect of solar activity and volcanic eruption on solar radiation is combined in the one index called the effective solar radiation. In order to evaluate the effect of each single factor，experiments of 1000 years in time interval were designed. The combining effect of different forcing factors on the global surface air temperature were analyzed and compared with the reconstruction results based on climate data. The simulated climate change trend by these two models is by and large in line with the reconstructed climate change. There were the following climate times in the past 1000 years: warm during 1000～1300，corresponding to the Medieval Warm Period(MWP); cold during 1300～1850，corresponding to the Little Ice Age(LIA); and warming since 1860.Especially for the maunder minimum of 1670～1710，the modeling result highly agree with the reconstruction result. The sensitive experiments show that the roles of forcing factors are varying with periods and time scales. It is found that the major difference between the simulated and the reconstructed temperature series would occur in the MWP of 1000～1300，significantly representing in phase and amplitude. In general，the simulated temperature anomalies are less than the reconstructed temperature anomalies. At the millennium scale，solar activity，volcanic eruption，and greenhouse gas are important for the global temperature change. On the other hand，for the past 100 years，the change of greenhouse gas had more contribution to the global temperature change. It is also found that the decrease of CO2 concentration would produce about 0.1℃ cooling in the LIA，accounting for one third of the temperature amplitude. The comparison of the results given by these two models shows that although they can be used to predict the climate trend，model MPM­2 seems to be more efficient and convenient than model ECHO­G in long range experiments.