A membership function consists of shapes comprised of its own linguistic terms. The most commonly used ones are triangular, trapezoidal, or parabolic in shape. The present study is based on triangular membership functions. Each linguistic term has a particular interval value. In classical fuzzy, these intervals are determined through a number of operations. However, experiences are emphasized in the system based on expert opinion. In gene-fuzzy, starting and ending values of these intervals are determined for each membership function following the stages of genetic algorithm. Algorithm is inserted into the cycle until the optimum result is found. Determined intervals are transferred to the fuzzy logic system. This method is assumed to yield a better result than classical fuzzy, which is based on simple mathematical calculations, and expert fuzzy, which is dominated by experiences.