Abstract: The quick and precise implementation of transient stability assessment（TSA） is of great significance for the safe operation of power system.In recent years,the rapid development of deep learning techniques has become the effective measures to deal with this issue.However,the deep learning models based on neural networks have some drawbacks including difficulty in parameter regulation,long training time and big demand of samples.In this paper,we establish a transient stability assessment model for power system based on deep forest.Some physical characteristics at the fault clearing moment are selected as the input features,and the transient stability state of a system is considered as the output result.The simulations on New England 39-bus system show that,compared with the deep neutral network,the proposed method has advantages in simple parameter setting,rapid training speed,moreover,it can effectively avoid over-fitting and has a good generalization ability even when the number of training samples is small.