The IMC controller based on RBF adaptive neural network is designed by the capacity of nonlinear approach of RBF neural network。 The problem of nonlinear and uncertainty bound is solute。 The weigh coefficient that is difference between expectation output of internal model and output of the object is designed。 The recurrence learn algorithm is used to carry out RBF neural network cluster center by SOFM neural network， and recurrence least square is used to fix on weigh of RBF neural network The simulation results show that new method leads to improve dynamic performance and robustness。
NETL
