In this paper we present a methodology for developing dynamical models for continuous digesters using input-output data. The methodology uses the Radial Basis Function (RBF) neural network architecture, which is continuously increasing its popularity in solving system identification problems, due its simple network structure and the short training times it employs. The produced dynamic RBF network model can be utilized to predict the future behavior of the process, analyze the dynamics of the digester and control the process through a Model Predictive Control (MPC) scheme.