In real-life situation, complexity occurs when the system is not understandable. The size, shape, position, or color of a system can cause a phenomenon. These complexities are classified in fuzzy sets. The fuzzy sets role is significant when applied to complex phenomena, which is not easily described by traditional mathematics. Fuzzy regression may be more appropriate tool than the traditional regression analysis. In fuzzy regression, it is assumed that the system’s structure is ambiguous or complex.A brief review of the linear programming models in fuzzy linear regression is given. Symmetric, asymmetric, trapezoidal and bell shaped membership functions are investigated. In addition, the different linear programming models are summarized, and a new model is presented for better results. Moreover, the new model is compared with other two models that are Tanaka and Hojati models.