The Mulan Weichang National Nature Reserve, located in northern Hebei province, adjacent to southeastern Inner Mongolia, is an important ecological barrier and nature reservoir for Beijing-Tianjin-Hebei region. Cognizing the historical climate change of this area will benefit the ecological protection and management of North China Plain. As an essential research technique for the climate variability and predictability (CLIVARl), dendroclimatology tackles problems of present and past climate with information of dated tree-rings. According to the theories of dendroclimatology, we developed two-group standard chronologies (STD), residual chronologies (RES) and autoregressive chronologies (ARS) based on tree-ring widths of Chinese pine (Pinus tabuliformis) and Yeddo spruce (Picea jezoenssis var. komarovii) from two sampling points in the nature reserve and all the chronologies were significantly correlated with each other. The correlation analyses between chronologies and climate factors showed that the radial growth of both arbor species was closely related to summer temperature and annual precipitation. Among all the calculations, STD of Chinese pine (STD_(pine)) and May-June mean temperature had the maximum correlation coefficient (r=-0.654, P<0.01). Additionally, we investigated the relationship between chronologies with ecoclimatic indices. The analysis suggested that tree-ring growth was highly influenced by humidity index and aridity index. Using regression analysis, the mean temperature of May to June, the temperature of May and annual aridity index from 1909 to 2008 were reconstructed on the basis of STD_(pine). After calibration, the explained variances of the three reconstructions value were 48.1%, 39.9% and 33.8% respectively. The trend of change of each estimated series was consistent with the observed data. In addition, the estimated and observed aridity indexes were both better correlated with temperature than precipitation, indicating that the contribution of temperature to the aridity of the study area was larger than precipitation. During the entire reconstructed series, the most distinctively warm period occurred between the 1910s and 1930s. Soon afterwards, the temperature had a rapid fall and reached the lowest point in 1935. After that, it gradually increased and lasted for the next five years, then began to decrease again. Although there was a slight rise from 1957 to 1968, the series showed a moderate descending tendency for the next 40 years. The temperature has increased rapidly since the middle 1990s, showing a warming period. There were four xerothermic periods and three clammy periods in the past one hundred years. The intervals above average of each reconstructed series comprised 1909-1929, 1939-1948, 1980-1989 and 1997-2008, while the below ones were consisted of 1930-1938, 1949-1979 and 1973-1979. A significant cycle of quasi 21[CD*2]year was also detected in the series. Moreover, spatial correlation analyses revealed that the reconstructions were regionally representative. The estimated May-June mean temperature and May temperature can be well comparable with previous research of surrounding areas, as well as the historical drought events.