Abstract： Based on the data of Phyllostachys heterocycla cv. pubescens in Jianou，Fujian ，the dry weight and energy of bamboo various parts were determined，their bi omass models were built，and the artificial neural network was applied to estima te the energy of the different parts of Phyllostachys heterocycla cv. pube scens forest．The results showed that the energy of stem，branch and leaf，und er9.2295ground part were 4.23225×108, 9.2295×107 and 1.7643×108 kJ hm -2，and their distributive proportion of energy were 61.32%，13.11% and 25.57%, respectively．The mean simulative accuracy of artificial neural models o f stem，branch and leaf，underground part were 87.88%，82.99%and 82．95% respect ively，and their predictive accuracy were 87.13%，81.32% and 81.86%, respectivel y．This paper provids a scientific basis in revealing the latent capacity of Phyllost achys heterocycla cv. pubescens forest． Fig 4, Tab 7, Ref 19