In this paper, a kind of traffic prediction a nd congestion control policy based on FNN (fuzzy neural network) is proposed for ATM (asynchronous transfer mode). Congestion control is one of the key problems in high-speed networks, such as ATM. Conventional traffic prediction method fo r congestion control using BPN (back propagation neural network) has suffered fr om long convergence time and dissatisfying precision, and it is not effective. T he fuzzy neural network scheme presented in this paper can solve these limitatio ns satisfactorily for its good capability of processing inaccurate information a nd learning. Finally, the performance of the scheme based on BPN is compared wit h the scheme based on FNN using simulations. The results show that the FNN schem e is effective.