The objective of this project is to investigate the feasibility of using neural networks to classify industrial parts according to process classes defined by Sundstrand Pacific Aerospace Pte. Ltd. The study is necessary to support the implementation of an automatic retrieval and storage system integrated with CATCH, a Computer-Aided Tolerance CHarting program currently used by Sundstrand. A successful implementation will lead to increase of productivity in the process planning procedure for parts and may also lead the way in standardising the process planning procedure. Master of Engineering (MPE)