Based on the climatic data collected from 1981 to 1990 in the actual distribution region of Larix gmelinii, Betula platyphylla, Picea koraiensis in Northeast China, 12 climatic factors,which affect the 3 species,were extracted by using the theory and methods of hierarchical clustering based on fuzzy proximity relations and clustering fusion technology. A random mathematical model for predicting the distribution of tree species was constructed by using the rigorous theory and methods of statistical analysis and data processing, and the corresponding algorithm was studied. The MATLAB was used to design and run program under ArGIS9. 3 interface,which produced predicting maps for the optimal adaptable distribution,the intermediate adaptable distribution and the general adaptable distribution of the three tree species in Northeast China from 2041 to 2050. By analyzing these maps, the distributions of the three tree species would drift to the north under climate change, L. gmelinii and B. platyphylla would drift to the northwest,and P. koraiensis would drift to the northeast. Furthermore,climate change was found to be the main impact factor in regulating the predicted distributions of the three tree species.