This paper presents the logic of a knowledge-based system for turning tool selection. The selection philosophy is based on machining performance and the system uses information regarding tools and cutting data from ‘approved’ operations which have been proved on machine tools following a specific testing procedure. For any new operation, rules are used to identify its level of similarity to previously performed, approved operations. The similarity criteria are based on metal cutting theory and practical engineering knowledge and incorporate considerations in relation to the component and cutting profile geometry, material type and operation type as well as tool and insert characteristics. The goal is to identify similar approved operations, retrieve the corresponding tool and cutting data and sort them in order of preference. A key function of the system is that according to the level of similarity, the retrieved information is either used as it is or is automatically modified to suit the new operation. The main benefits from using the system are improved engineering consistency in the decision making for selecting tools and cutting conditions, improved utilization of tools and definition of efficient machining conditions.