In the paper, learning system of manufacturing knowledge acquisition in the range of tool selection to machining operation was presented. Currently often and often research concerning machine learning is developed. Machine learning includes issues of system designing which improves its operations, along with analysis of experience represented by file of learning examples. Recently, machine learning methods have been applied successfully in many practical problems and are becoming a part of advanced information systems - in particular concerning knowledge discovery in databases, and the so-called data mining. Learning system acquires knowledge by using the method of decision trees induction. This method allows approximation of classification functions of discrete output values relating to certain terms, decision classes. The system learns to select tools on the basis of the ones chosen out of the database by process engineering experts for machining operations. In the method of decision trees induction a decision tree is created. This tree allows classification of the whole learning examples file into homogeneous classes. On the basis of the decision tree, decision rules are created. Next, these rules are used in expert system for selection of tools from outside of the file of learning examples.