Abstract: As high spatial resolution remote sensing images and LiDAR(light detection and ranging)technology applied in forestry to extract tree crown and height automatically, new models are needed to predict forest stand information. Based on the investigation of 178 sample plots distributed in Beijing, the most common used crown diameter(C)-diameter at breast height(D) models and tree height(H)-D models were chosen to predict D from C and H for thirteen common tree species in Beijing. The results showed that not all tree species'C and H had close relationship with D. Tree species of Ailanthus altissima, Cedrus deodara, Koelreuteria paniculata and Populus canadensis had high correlation of D-C and D-H equation, with determination coefficient (R2) value higher than 0.7 and 0.5, and above 0.8 by combination C and H to estimate D. But the C of aspen, China savin, Oriental white oak, Chinese pine and the H of walnut, willow, Chinese scholar tree, torch tree, Chinese pine were poor correlated with D, both with a R2 value lower than 0郾3. Regional site conditions and stand variables (site index, stand age and density) were suggested to join with C and H to improve the D predicting accuracy in the next step work. This method can be used in estimating D and other stand information with C and H automatically extracted by modern remote sensing technology, which can realize quick and economical forest resource investigation and update.