MACHINE TOOLS MUST NOT ONLY BE CAPABLE OF PRODUCING PARTS WITH GREATER GEOMETRICAL ACCURACY, BUT ALSO MORE CONSISTENTLY FROM ONE TO THE NEXT OVER LONG PRODUCTION RUNS. THERMAL RELATED ERRORS CONTINUES TO THWART HIGH PRECISION MACHINING APPLICATIONS. THE FEASIBILITY OF NEURAL NETWORKS FOR PREDICTING AND COMPENSATING FOR MACHINE TOOL THERMAL ERRORS WAS INVESTIGATED. THE STUDY FOCUSED ON A TURNING CENTER WHERE THE NEURAL NETWORK SHOWED GOOD CORRELATION TO MEASURED DISTORTION OVER A WIDE RANGE OF THERMAL CONDITIONS- -BOTH INTERNAL AND EXTERNAL THERMAL INFLUENCES. THE NEURAL NETWORK PREDICTED THE MAGNITUDE, TREND AND PATTERN OF THE THERMAL ERROR TO WITHIN 2%.