Autonomic learning(AL) refers to an inner mecha- nism of self-directed learning integrated by learner's attitude, capability and learning strategy. AL usually means active, self-conscious and independent learning, which is opposite to the type of passive, mechanical or receptive learning. AL has always been a hot issue of machine learning research. In this paper, based on the theory of Dynamic Fuzzy Logic(DFL), autonomic learning model and algorithm are developed, which provide a theoretical basis for the people to solve this type of problem. Simulation results illustrate the effi- ciency of this autonomic learning method.