We propose a novel approach to automated fuzzy rule base creation based on underlying observational data. The core of this method lies in adding information to particular fuzzy rule in the form of attached weight given as a value extracted from a relational data model.