Abstract: Few studies focus on the growth ring of broadleaf trees in Taiwan at present, especially lack of researches in modeling growth rate of natural broadleaf forest. This study thus aims at studying the growth ring of typical warm temperate broadleaf forest in central Taiwan. The 18 long-term forest dynamics plots in shenmuhsi reserved forest were established since 2003. Trees with DBH 6 cm in 6 plots of 13--18 were coring at breast height. 448 cores in total from 39 species 33 genera 22 families were obtained based on the stand characteristics. Microscope and CCD were used for tree-ring analysis. The transition point between successive annual rings was ditermined and classified visually, and then was used as the basis for visual determination and calculation of growth ring. We measured ring width using Velmax system. The crown-class model (CCM), PAI regression model, and mixed model were used to describe a crown trend and to analyze the age structures of species, which could be served as the references for the stand management on most widely distributed Machiluscastanopsis and Quercus with largest area in Taiwan. Regarding the character of growth ring, 79% samples are diffuse-porous species, 18% semi-ring-porous and 3% ring-porous. Average growth rate for all species is 3.53 mm/a. The significant vertical difference happened in canopy level. The average growth rate is 4.47 mm/a in overstory (upper) trees, which is the fast growing one, 2.95 mm/a in overstory (main) trees, and 2.92 mm/a in understory trees. PAI model denotes that the ring width linearly increases over time in most of the species. With 3 models of crown-class, PAI and mixed the estimated average age was 55, 60 and 59 respectively. The use of mixed model could lower error percentage than using PAI mode and crown-class model. The distribution of age group from CCM, PAI and mixed model revealed that this area is full of shade-tolerant seedlings and shrubs, and then gradually typical multi-layer forest is formed. The trees by age could be divided into two groups. The first group is mostly shade-tolerant seedlings originated at the late 20th century, and the second group formed an independent peak in 1900. Invisibility of growth ring caused the difficulty on classification. Irregularity of ring width resulted in the no way to determine age. PAI model needs large sample data of tree cores from different tree ages to do regression for being representative. Other problems still remain although the mixed model could lower estimation error of tree age. Therefore, it is necessary to further discuss and analyze the information and methodologies for the sustainable management and development of forest ecosystem. The study provides a sample of radial growth and age structure of warm temperate broadleaf stand for better understanding of forest management in this area.