Abstract: The difficulty of modeling an extraction/separation process of rare earth metal to implement its automatic control is mainly lacking an exact mathematical description. An intelligent optimal control method is therefore proposed using case-based reasoning (CBR) technique to preset relevant process parameters, such as the flowrates of extractant, feed liquid and scrubing solution, with the aim of controlling optimally the purity of products from the outlets at both ends to up to the target indices. The extraction/separation process can thus be understood in depth from a lot of historical data accumulated in previous cases so as to build a case database summarizing the typical operation conditions. Discusses the case retrieval, adjustment, revision and store in CBR process. The method proposed has been applied to an HAB extraction/separation process control for yttrium (Y), and its effectiveness was proved evident.