The most important prerequisite for a thorough software test is the determination of relevant test cases. The classification-tree method suggested in this paper supports the systematic determination and description of test cases. It is based on the idea of partition testing. By means of the classification-tree method, the input domain of a test object is regarded under various aspects assessed as relevant by the tester. For each aspect, disjoint and complete classifications are formed. Classes resulting from these classifications may be further classified—even recursively. Test cases are formed by combining classes of different classifications. The stepwise partition of the input domain by means of classifications is represented graphically in the form of a tree. This tree is subsequently used to form a combination table in which the test cases are marked. Extensions of the graphical notation and tool support allow the use of the method even for extensive test problems.