Abstarct: System Identification, as it is known, is a field of control theory concerned with the fitting of models to given sets of signals or data. Most existing methods for system identification make use of an input-output frxamework where the input signals, u(t) to the system, and the output signals, y(t), from the system are observed data. Hybrid systems have attracted increasing attention in the control community during the last decade. This interest is mainly due to their ability to capture the mixed continuous and discrete dynamics of many real systems. Hybrid Systems consist of continuous time and/or discrete time processes interfaced with some logical or decision making process. In this thesis, we focus on the identification of hybrid systems. An on-line system identification method, namely evolutionary potential fuzzy clustering approach is developed with some useful modifications for hybrid systems. The effectiveness of the proposed identification method is tested and evaluated on some benchmark hybrid systems.