THIS PAPER INTRODUCES A PROJECT INVOLVING THE DEVELOPMENT OF NEURAL NETWORK (CONNECTIONIST) EXPERT SYSTEMS FOR MACHINING QUALITY ANALYSIS AND DIAGNOSIS IN A CIM ENVIRONMENT. THE DEVELOPMENT OF NEURAL NETWORK EXPERT SYSTEM AND THE COMPARISON TO RULE-BASED EXPERT SYSTEMS ARE DESCRIBED. THE MACHINING QUALITY ANALYSIS AND DIAGNOSTIC PROBLEMS AND THE MACHINING ERROR DIAGNOSTIC MODEL ARE PRESENTED. A STRUCTURE FRAMEWORK OF NEURAL NETWORK EXPERT SYSTEMS IS PROPOSED WHICH INCLUDES KNOWLEDGE BASE, INFERENCE ENGINE, LEARNING MODULE AND AN EXPLANATION FACILITY. A PROTOTYPE SYSTEM FOR BOX-TYPE PART QUALITY ANALYSIS AND DIAGNOSIS IS GIVEN. PROBLEMS ENCOUNTERED DURING THE DEVELOPMENT ARE ALSO DISCUSSED.