This paper addresses a new approach for automatically extraction of the type-2 fuzzy rules from input-output data. In this approach, the structure identification and parameter optimization are carried out automatically without any assumption about the structure of the data, which is capable of finding the optimal number of the rules with an acceptable accuracy. The fuzzy modeling algorithm proposed in this paper is composed of three significant modules: (1) generate an initial rule-base, (2) tune the rule-base, (3) construct a new rule and add to rule-base. First, an initial two-rule fuzzy model (type-1) is generated, and then, this rule base is transformed to interval type-2 fuzzy rule base. According to the system output error, another type-2 rule is added to rule base. This procedure is continued until the stop criteria are satisfied. Four examples are provided to evaluate the performance of the proposed modeling approach.