Abstract:The research on forest self-thinning has been developed greatly, and a lot oftheoretical or empirical formulations were proposed. In this paper a method of artificialneural network is introduced, and a model of three-level feedtoward back-propagationneural network to simulate self-thinning law of Pinus massoniana is established for the firsttime. The result of simulation shows that artificial neural network model can fit the observeddata very well, and it is very useful in practice, which will enrich the research method of forest self-thinning.