This paper proposes a method which identifies unknown systems using fuzzy-rule based models(fuzzy models) when the input-output pairs of the system are given. It searches fuzzy models by genetic algorithms based on the given input-output pairs. The method finds all parameters of fuzzy models : the number and the position of the fuzzy sets of each input and the rule base. We encode only the fuzzy partitions of inputs into chromosomes, and then generate fuzzy rules from the encoded fuzzy partitions and the given data. We evaluate the performance with 3 functions. The experiments show that the proposed method properly locates the fuzzy sets on the input domains and generates the fuzzy rules approximating the given data. Keywords : Identification, Fuzzy Model, Genetic Algorithms, Least Squares Estimate. 1 Introduction Recently, the modeling methods using fuzzy rules and fuzzy reasoning have been widely used in many application areas and showed good results.