Fuzzy modelling is simultaneously an extension of existing modelling approaches and a negation of one of their main aspects, the crispness of their definition. As a digital equivalent of analogue sketching it allows designers to register and manipulate imprecise and uncertain information. In the framework of design representations fuzzy modelling supports the development of conceptual design models characterized by flexible definition and interaction. The main advantages of such models are fluency, abstraction and continuity, at a level similar to that of analogue techniques. In addition to that they offer the possibility of local autonomy, i.e. segmentation of a representation into self-regulating and cooperating components. Three alternative forms of fuzzy modelling are proposed: (1) Canonical objects with tolerances, (2) objects described by minimal and maximal values, and (3) point sets comprising discrete, autonomous particles that describe the object by their spatial or structural relationships.