If the fuzzy returns of securities are some special variables such as triangular fuzzy variables or trapezoidal fuzzy variables, the models (3) and (4) can be solved through analytical way. However, when the membership functions of fuzzy returns take more complex forms, generally, it is difficult to analytically solve the new models (3) and (4). To provide a general solution to the new models, we design a hybrid intelligent algorithm integrating genetic algorithm (GA) and fuzzy simulation. GAs were proposed by Holland [8] in 1975, and have been well developed since then, such as Koza [9]. Particularly, Buckley and Hayashi [3], Liu and Iwamura [12], Liu [13–15], and Buckley and Feuring [4] designed a spectrum of GAs for solving fuzzy programming models. As for fuzzy simulation, Liu introduced the technique in detail in book [17]. Broadly speaking, in the proposed hybrid intelligent algorithm, the technique of fuzzy simulation is applied to compute the credibility or to find the a-return _f first. Then, fuzzy simulation and GA are integrated for solving the fuzzy models